JP5566674B2 - Interpretation report creation support system - Google Patents

Description

  The present invention relates to an interpretation report prepared as a medical observation report for a patient image taken by a medical imaging apparatus such as an X-ray computed tomography apparatus (CT) or a magnetic resonance imaging apparatus (MRI). The present invention relates to an interpretation report creation support system that can be referred to, and more particularly, to an interpretation report creation support system that can refer to correction contents of an interpretation report created or modified by a plurality of interpretation doctors.

  Conventionally, a patient's image captured by a medical imaging device has been interpreted (screening the findings in the examination image), and a report (interpretation report) summarizing the results of the interpretation has been prepared. Created.

  In creating an interpretation report, two or more interpretation doctors may interpret the image of the examination. For example, when two interpreting doctors perform interpretation, an interpreting report created by one interpreting doctor (hereinafter referred to as “primary interpreting doctor”) is read by another interpreting doctor (hereinafter referred to as “secondary interpreting doctor”). The contents may be corrected. In such a case, an interpretation report creation support system that enables the primary interpretation doctor to confirm the content corrected by the secondary interpretation doctor is known as the prior art.

  In the conventional interpretation report creation support system, if the contents of the created report (character string) are changed by an interpretation doctor (for example, secondary interpretation doctor) different from the primary interpretation doctor who created the report, the report is corrected. This is communicated to the primary interpreting physician. At this time, the primary interpreting doctor can read the contents of the report changed to another interpreting doctor and confirm what correction has been made.

  On the other hand, there is known an interpretation report creation support system capable of extracting another interpretation report that is closely related to a certain interpretation report as disclosed in Patent Document 1.

  In the interpretation report creation support system, morphological analysis is performed on sentences included in the report, and a description unit including a part word, an event word, and a presence / absence word is created in advance from the finding sentence, thereby semantically converting the finding sentence. Analyze and structure. According to the interpretation report creation support system, it is possible to extract reports that are semantically related by searching for reports based on the description units.

JP 2009-129108 A

  When two or more interpreting doctors interpret the image of the examination, minor corrections such as correction of typographical errors and conversion mistakes, and medically important corrections such as finding corrections, etc. Are mixed. In such a case, the conventional interpretation report creation support system reflects and displays these corrections without distinction.

  The primary interpretation doctor confirms which part of the corrected report is modified and how based on the display by the interpretation report creation support system. Therefore, if a lot of minor corrections are included, an important correction point may be overlooked.

  The present invention solves the above problem, and when corrections are made to an interpretation report, by analyzing the correction contents, only the correction contents including important corrections are transmitted to the primary interpretation doctor. The purpose is to provide an interpretation report creation support system.

In order to achieve the above object, the invention described in claim 1 is an interpretation report creation support system including a report management unit that identifies and stores an interpretation report of the same examination for each version. A difference extraction unit that extracts different descriptions generated by correction as correction differences for an interpretation report with different corrections, and specifies a correction method based on the different description modes of the correction difference, and a character string of the correction difference by analyzing the mean content and aspects of the sentence structure, the correction difference, the the modified content extraction unit that specifies whether to correspond to any of a plurality of types including finding sentence, the identified correction method of the type based on the combination of, by referring to the table showing the classification of modified type by a combination of the type of modified difference and correction method including pre-noise modified type containing a modified findings sentence Or to identify the Re, in accordance with the identified said modified type, characterized by comprising a determining transmission determination unit that determines whether to transmit the modified difference.

  According to the present invention, it is possible to determine the correction type from the content of correction applied to the interpretation report and determine whether or not to transmit the correction content to the primary interpretation doctor based on the correction type. . At this time, by setting so that only the correction type including the important correction is transmitted to the primary interpretation doctor, it is possible to suppress the transmission to the primary interpretation doctor regarding the correction for the minor defect and transmit only the important correction. It becomes possible. As a result, it is possible to reduce the burden of work for the primary interpretation doctor to confirm the correction contents.

It is a system block diagram of an interpretation report creation support system according to the present invention. It is an example of the input information and output information of a character information difference extraction part. It is an example of the input information and output information of an image information difference extraction part. It is an example at the time of dividing the correction difference of a character string into each section. It is a figure for demonstrating the method of discriminating a correction method from the correction difference of a character string. It is a flowchart of the process which classify | categorizes the correction difference of a character string into the kind of correction difference. It is an example of the data structure of a medical term dictionary. It is an example of the data structure of a misprint and an abbreviation list dictionary. It is an example of the data structure of the structured dictionary used by creation of the description unit by structured processing. It is a figure for demonstrating a structuring process. It is an example of the data structure of a character string difference management table. It is an example of the data structure of a correction classification determination table. It is an example of the data of the correction type determination table in the case where the correction type is incorrect correction. It is an example of the data of the correction type determination table in the case of correction of the correction type. It is an example of the data of the correction type determination table when the correction type is correction of a finding sentence. It is an example of the data structure of a transmission determination table. It is an example of the display of the correction content regarding the correction difference of a character string. It is an example of the display of the correction regarding the correction difference of the combination of an image and a character string.

(First embodiment)
An interpretation report creation support system according to the first embodiment will be described with reference to FIG. The interpretation report creation support system according to the present embodiment is an interpretation report creation support system including a report management unit 10 that identifies and stores an interpretation report for each version, and further includes a difference extraction unit 20 and a correction. A type determination unit 30, a transmission determination unit 40, a transmission content storage unit 50, and a correction content display method selection unit 60 are provided. Each component may be built in different computers, and these computers may be connected via a network, or all the components may be built in one computer. The configuration shown in FIG. The combination of each component is not limited as long as the element is included.

  The function of each component and the content of processing will be described below.

  The report management unit 10 includes a report storage unit 11 that stores the created interpretation report for each version, and a version management unit 12 that selectively extracts a predetermined version of the interpretation report related to a predetermined inspection.

  The report storage unit 11 stores and manages the created interpretation report together with information on the interpretation doctor who created the interpretation report. Interpretation report is a result of interpretation of patient information such as patient ID and patient name, examination information such as examination name and examination site, X-rays taken in examination, CT image, and findings or diagnosis. And key images attached as reference information. The report storage unit 11 can extract the interpretation report corresponding to the search key using the description of the interpretation doctor information, patient information, examination information, findings or diagnosis as described above as a search key.

  Further, when the description of the already-stored interpretation report is updated, the report storage unit 11 stores the updated interpretation report so that it can be identified as a version different from the interpretation report before the update. Specifically, the report storage unit 11 stores the updated image interpretation report as data different from the image interpretation report before the update, so that the image interpretation report before the update and the updated image interpretation report are stored in different versions. Identify as. At this time, the report storage unit 11 stores the newly created version together with the information of the interpretation doctor who created the version. Each time the interpretation report is updated, the report storage unit 11 stores the interpretation report as a new version.

  The version management unit 12 receives a search key (for example, patient information, examination information, findings, diagnosis, etc.) described above as a condition for extracting an interpretation report and a version designation as input, and from the report storage unit 11, An interpretation report corresponding to the specified inspection and version is extracted. Thereby, for example, for the interpretation report of the same examination, it is possible to selectively extract the latest version of the interpretation report and the interpretation report of the version immediately before the latest version.

  In addition, the version management unit 12 is preferably configured to extract the version created or modified by the interpretation doctor by designating the information of the interpretation doctor as a condition for extracting the interpretation report.

  The operations of the report storage unit 11 and the version management unit 12 will be described with specific examples. For example, assuming that the interpretation report created by the interpretation doctor α is the first version, when the interpretation report is corrected (updated) by the interpretation doctor β, the corrected interpretation report is stored in the report storage unit 11 as the second version. The At this time, both the first version of the interpretation report created by the interpretation doctor α and the second version of the interpretation report corrected by the interpretation doctor β are selectively extracted from the report storage unit 11 via the version management unit 12. Is possible.

  The difference extraction unit 20 receives two different interpretation reports as inputs, compares the descriptions of the two interpretation reports, extracts different descriptions as one or more correction differences, and includes a comparison including the one or more correction differences. Information on the result (hereinafter referred to as “comparison result”) is output to the correction type determination unit 30. At this time, for example, by comparing descriptions between two different versions of an interpretation report of the same examination, it is possible to extract a correction applied to the interpretation report as a correction difference. In the following description, it is assumed that two different versions of the interpretation report of the same examination are compared, and at this time, the old version is called “old version report” and the new version is called “new version report”. The old version report corresponds to the first version, and the new version report corresponds to the second version.

  The difference extraction unit 20 includes a character information difference extraction unit 21 and an image information difference extraction unit 22.

  The character information difference extraction unit 21 compares a character string described in the new version report (for example, a character string described as a finding or diagnosis) with a character string described in the old version report, and only the new version report has a corrected difference. Either or both of the correction difference of the included character string and the correction difference of the character string included only in the old version report are extracted. Hereinafter, individual correction differences extracted by the character information difference extraction unit 21 may be referred to as “character string correction differences”. The character information difference extraction unit 21 outputs the comparison result information (hereinafter referred to as “character string comparison result”) including the correction difference of one or more extracted character strings to the correction type determination unit 30 as a comparison result. .

  The processing of the character information difference extraction unit 21 will be specifically described with reference to FIG. FIG. 2 is an example of input information and output information of the character information difference extraction unit 21. The character information difference extraction unit 21 receives the new version report and the old version report as input, and compares the character string described in the new version report with the character string described in the old version report. The character string comparison method is generally known, such as the diff command algorithm.

Based on the result of the character string comparison, the character information difference extraction unit 21 corrects the character string in accordance with the character string comparison result output rules shown in <Rule A-1> to <Rule A-4> shown below. Outputs the “character string comparison result” including the difference. An output example of information (character string comparison result) output by the character information difference extraction unit 21 is shown as “character string comparison result” in FIG.
<Rule A-1> Character strings that are not changed between the new version report and the old version report are output with “” (half-width space) added to the beginning of the line.
<Rule A-2> Line feed is output as “\ n”.
<Rule A-3> Character strings deleted when the old version report is revised to the new version report are output with “-” added to the beginning of the line.
<Rule A-4> The character string added when the old version report is revised to the new version report is output with “+” added to the beginning of the line.
Note that the character string corresponding to <Rule A-3> and <Rule A-4> indicates a correction difference of the character string.

  First, the image information difference extraction unit 22 uses information (hereinafter referred to as “combination information”) that combines the image associated with the character string described in the interpretation report and the character string for each of the new version report and the old version report. As a list (hereinafter referred to as “combination information list”). At this time, for example, the image information difference extraction unit 22 identifies an image from a link associated as a tag with the character string and extracts combination information.

  Next, the image information difference extraction unit 22 compares the list of combination information extracted from the new version report and the old version report, and as the correction difference, the correction difference of the combination information included only in the new version report and the combination included only in the old version report Either or both of the information correction differences are extracted. In the following, individual correction differences extracted by the image information difference extraction unit 22 may be referred to as “combination information correction differences”. The image information difference extraction unit 22 outputs the comparison result information (hereinafter referred to as “combination information comparison result”) including the correction difference of the extracted combination information at one or more locations to the correction type determination unit 30 as the comparison result. .

  The processing of the image information difference extraction unit 22 will be specifically described with reference to FIG. FIG. 3 is an example of input information and output information of the image information difference extraction unit 22.

  The description in the old version report in FIG. 3 is associated with images with identifiers of IMG1, IMG2, and IMG3. At this time, for example, the combination information indicating the image corresponding to IMG1 and the character string “consolidation of the upper right lobe of the right lung” associated with the image is (IMG1, “consolidation of the upper lobe of the right lung”) Are managed in various formats. The image information difference extraction unit 22 extracts a list of combination information in the old version report as a set P in the following format.

  P = {(IMG1, "Right upper lobe consolidation"), (IMG2, "Bronchial translucency image"), (IMG3, "")}

  Further, the description in the new version report in FIG. 3 is associated with images with identifiers of IMG1 and IMG2. The image information difference extraction unit 22 extracts a list of combination information in the new version report as a set A in the following format.

  A = {(IMG1, “Right upper lobe consolidation”), (IMG2, “Cavity inside”)}

  The image information difference extraction unit 22 calculates the difference set between the set P and the set A as follows, and extracts the correction difference of the combination information. For example, combination information included only in the old version report (combination information deleted between the old version report and the new version report) is obtained as the difference set P / A as follows.

  P / A = {(IMG2, "bronchial transillumination inside"), (IMG3, "")}

  Further, combination information included only in the new version report (combination information added between the old version report and the new version report) is obtained as a difference set A / P as follows.

  A / P = {(IMG2, “Cavity inside”)}

  Next, the image information difference extraction unit 22 obtains the union (P / A) ∪ (A / P) of P / A and A / P. Thereby, both the combination information included only in the old version report and the combination information included only in the new version report are extracted. The union (P / A) Ａ (A / P) of P / A and A / P is obtained as follows in the example of FIG.

  (P / A) ∪ (A / P) = {(IMG2, “Bronchial translucency image”), (IMG2, “Cavity inside”), (IMG3, “”)}

  When (P / A) ∪ (A / P) is an empty set Φ, set P = set A. In this case, the image information difference extraction unit 22 has not extracted the correction difference of the combination information between the new version report and the old version report.

  In the following description, it is assumed that the image information difference extraction unit 22 outputs (P / A) ∪ (A / P) as a comparison result of combination information.

  In the above description, the identifier of the image and the character string associated with the image are used as constituent elements of the combination information. However, the present invention is not limited to these constituent elements. For example, it may be configured only by an image identifier, or other components such as gradation and annotation may be added. When the image identifier is constituted only by the image identifier, the image identifier may be extracted from the link associated with the character string as a tag and used as combination information. In addition, when adding gradation or annotation as a constituent element, it is possible to refer to image data from a link associated with the tag, extract gradation or annotation information included in the image data, and use it as a constituent element of combination information. good. Note that the method for calculating the comparison result of the combination information does not change.

  In the above description, all the components constituting the combination information are to be compared, but only some components may be compared. For example, when only the identifiers of the images are to be compared, the combination information comparison result (P / A) ∪ (A / P) in FIG. 3 is (P / A) ∪ (A / P) = {(IMG3 , “”)}.

  The correction type determination unit 30 receives the comparison result from the difference extraction unit 20, specifies a correction method for each of one or more correction differences included in the comparison result, further analyzes the correction difference, and performs the correction Based on the analysis result of the difference, the classification of the correction difference (this classification is “classification for each type of correction difference”, hereinafter referred to as “class of difference”) is specified.

  The correction type determination unit 30 identifies a correction method for each of the correction differences extracted by the difference extraction unit 20 by detecting whether the correction difference is included in the new version report or the old version report. In addition, the correction type determination unit 30 analyzes each correction difference and classifies it into a difference class. In this classification, processing differs depending on whether the correction difference of the character string is classified or the correction difference of the combination information is classified. In addition, when classifying correction differences of character strings into a class of differences, classification based on the semantic content of correction differences of character strings and classification based on the form of sentence structure of correction differences of character strings are used. The details of the process for specifying the correction method and the details of the process for classifying into a difference class will be described later.

  In addition, the correction type determination unit 30 includes a dictionary storage unit 31 that stores dictionary data that is referred to when the correction differences of the character strings extracted by the character information difference extraction unit 21 are classified into a difference class.

  The dictionary storage unit 31 includes a first dictionary storage unit and a second dictionary storage unit (the first dictionary storage unit and the second dictionary storage unit are not shown). 7 and 8 show examples of dictionary data stored in the first dictionary storage unit. FIG. 9 shows an example of dictionary data stored in the second dictionary storage unit.

  In the first dictionary storage unit, a medical term dictionary (see FIG. 7) storing character strings indicating disease names and parts for each classification, and examples of character strings indicating abbreviations and misprints are collected and stored for each classification. Data of a list dictionary (see FIG. 8) is stored. The dictionary data stored in the first dictionary storage unit is referred to when character string correction differences are classified based on semantic content. The second dictionary storage unit stores data of a structured dictionary (see FIG. 9) in which character strings indicating part words, event words, and presence / absence words are stored for each of these classifications. The dictionary data stored in the second dictionary storage unit is referred to when the correction differences of character strings are classified based on the form of sentence structure. Details of the processing of the correction type determination unit 30 will be described below.

  First, the process for the correction difference of the character string by the correction type determination unit 30 will be described with reference to FIGS.

When receiving the character string comparison result from the character information difference extraction unit 21, the correction type determination unit 30 first identifies the synchronization point between the new version report and the old version report for the comparison result of the character string. Of the comparison results of the character strings, a portion corresponding to any of the following conditions is a synchronization point.
<Condition A-1> Position of “\ n” in the line beginning with “\ n” <Condition A-2> Beginning with “” (half-width space) and “.” ) The position of the first “.” (Read mark) in the line containing
However, when the synchronization point corresponding to <Condition A-1> and the synchronization point corresponding to <Condition A-2> are adjacent, the two adjacent synchronization points are processed as one synchronization point.

  After identifying the synchronization point, the modification type determination unit 30 divides the comparison result of the character strings at the synchronization point. Each unit divided at this time is called a section. FIG. 4 shows an example in which the character string comparison result output by the character information difference extraction unit 21 in the example of FIG. 2 is divided into sections. In FIG. 4, the character string comparison result is divided into sections SEC1, SEC2, and SEC3 by synchronization points P1, P2, and P3.

  The correction type determination unit 30 divides the character string comparison result into sections, and then specifies the correction method and classifies the correction differences of the character strings included in each section into a class of differences.

(Identification of correction method)
First, the correction type determination unit 30 specifies a correction method for each correction difference of the character string for each section. The correction method includes “replacement (addition)”, “replacement (deletion)”, “addition”, and “deletion”. Processing relating to the specification of the correction method of the correction type determination unit 30 will be described below with reference to FIG. FIG. 5 is a diagram for explaining a method of determining the correction method from the correction difference of the character string.

The correction type determination unit 30 determines whether the section satisfies all of the following conditions, and specifies a correction method for the correction difference of each character string included in the section based on the determination result.
<Condition B-1> There is one modification difference (line with “-” added to the beginning of the line) of the character string deleted when the old version report is revised to the new version report.
<Condition B-2> There is one modification difference (a line with “+” at the beginning of the line) added when the old version report is revised to the new version report.
<Condition B-3> The correction difference of the deleted character string and the correction difference of the added character string are adjacent to each other.

  When the section satisfies all of <Condition B-1>, <Condition B-2>, and <Condition B-3>, the modification type determination unit 30 corrects the difference between character strings included in the section to be determined. Among them, the correction method of the correction difference of the deleted character string is specified as “replacement (deletion)”, and the correction method of the correction difference of the added character string is specified as “replacement (addition)”. For example, the correction differences DIF21 and DIF22 of the character string included in the section SEC2 in FIG. 5 correspond to this case. In this case, the correction method of the character string correction difference DIF21 is replaced (deleted), and the correction method of the character string correction difference DIF22 is replaced (added).

  When the section does not satisfy at least one of <Condition B-1>, <Condition B-2>, and <Condition B-3>, the correction type determination unit 30 determines whether the character included in the determination target section. Among the correction differences of the column, the correction method of the correction difference of the deleted character string is specified as “deletion”, and the correction method of the correction difference of the added character string is specified as “addition”. For example, the character string correction differences DIF31 to DIF35 included in the section SEC3 in FIG. In this case, the correction method for the character string correction differences DIF31 to DIF33 and DIF35 is added, and the correction method for the character string correction difference DIF34 is deleted.

(Classification into difference classes)
Next, the modification type determination unit 30 classifies each modification difference of the character string into a difference class. The process of classifying the modification differences of the character strings into the class of differences by the modification type determination unit 30 will be described below with reference to FIG. FIG. 6 is a flowchart of a process for classifying correction differences of character strings into a class of correction differences. The modification type determination unit 30 classifies each modification difference of the character string into a difference class by performing the process shown in FIG.

(Step S1, Step S2)
First, the correction type determination unit 30 collates the correction difference of the character string to be processed with the medical term dictionary (see FIG. 7) stored in the first dictionary storage unit of the dictionary storage unit 31, thereby Analyzes character string modification differences based on semantic content. In addition, the correction difference of the character string is collated with an error or abbreviation list dictionary (see FIG. 8) stored in the first dictionary storage unit, so that the correction difference of the character string indicates an error or an abbreviation. It is analyzed whether it is. Specifically, the correction difference of the character string to be processed is compared with the dictionary data stored in the first dictionary storage unit, and the character string corresponding to the correction difference of the character string is registered in the dictionary data. Make sure that

(Step S3)
When the correction difference of the character string to be processed is registered in the dictionary data stored in the first dictionary storage unit (Y in step S2), the classification corresponding to the correction difference of the character string is specified, and the classification Classify as a class of differences based on 7 and FIG. 8 will be specifically described. For example, when the correction difference of the character string is “cervical spondylosis”, the classification is “disease name”, and the class of difference of “cervical spondylosis” is “disease name”. It becomes. Similarly, when the correction difference of the character string is “lung”, the classification is “part”, the corresponding difference type is “part”, and when the correction difference of the character string is “lacnar”, the classification is “erroneous”. The corresponding difference class is “erroneous”, and when the character string correction difference is “EC”, the classification is “abbreviation” and the corresponding difference class is “abbreviation”.

(Step S4)
When the modification difference of the character string to be processed is not registered in the dictionary data stored in the first dictionary storage unit (step S2, N), the modification type determination unit 30 performs morpheme on the modification difference of the character string. The analysis is performed, the correction difference of the character string is decomposed into morphemes, and the correction difference of the character string is analyzed based on the form of the sentence structure. Note that the morphological analysis technique is generally well-known as disclosed in, for example, “Information Processing, Vol. 41, No. 11, pp. 1208-1214, November 2000”, and thus is specific. The detailed explanation is omitted.

(Step S5)
First, the correction type determination unit 30 determines whether the correction difference of the character string to be processed includes an adjunct or adverb.

(Step S6)
When the modification difference of the character string to be processed does not include an adjunct or adverb (N in Step S5), the modification type determination unit 30 classifies the modification difference of the character string as “other words / phrases” as a class of difference.

(Step S7)
When the modification difference of the character string to be processed includes an adjunct or adverb (step S5, Y), the modification type determination unit 30 determines whether or not the modification difference of the character string is composed only of the appendix. .

(Step S8)
When the modification difference of the character string to be processed is composed only of an attached word (step S7, Y), the modification type determination unit 30 classifies the modification difference of the character string as an “adjunct word” as a type of difference. .

(Step S9, Step S10)
When the modification difference of the character string to be processed is not composed of only an appendix (step S7, N), the modification type determination unit 30 performs a structuring process on the modification difference of the character string and determines the description unit. Determine whether it can be created. Creation of description units by structuring processing will be described with reference to FIGS. FIG. 9 is an example of the data structure of the structured dictionary stored in the second dictionary storage unit of the dictionary storage unit 31 used in creating the description unit by the structuring process. FIG. 10 is a diagram for explaining the structuring process. Note that the description unit creation by the structuring process is described in Japanese Patent Application Laid-Open No. 2004-133830, and thus detailed description thereof is omitted.

  The modification type determination unit 30 confirms whether a description unit can be created from the modification difference of the character string to be processed.

  The processing relating to creation of the description unit will be specifically described with reference to FIG. The description unit is composed of a part word, an event word, and a semantic word. The modification type determination unit 30 collates each morpheme included in the modification difference of the character string to be processed with the structured dictionary stored in the second dictionary storage unit of the dictionary storage unit 31 to determine the part word wb, the event The word wj and the presence / absence word wu are extracted. In the case of the example of FIG. 10, the correction type determination unit 30 is “seen” from the correction difference DIF 51 of the character string as “left lung field” as the part word wb, “pleural effusion” as the event word wj, and presence / absence word wu. To extract. In this case, since the correction difference of the character string includes the part word wb, the event word wj, and the presence / absence word wu, the correction type determination unit 30 determines that the description unit ws can be created.

(Step S11)
When it is determined that the description unit ws can be created by performing the structuring process on the correction difference of the character string to be processed (step S10, Y), The correction difference of the character string is classified as a “finding sentence” as a kind of difference.

(Step S12)
The correction type determination unit 30 does not include at least one of the part word wb, the event word wj, and the presence / absence word wu in the correction difference of the character string to be processed, and cannot create the description unit ws. (Step S10, N), the correction difference of the character string is classified as “fragment sentence” as a class of difference. As described above, the correction method and the type of difference are specified for each correction difference of the character string.

  The correction type determination unit 30 stores the correction method and the difference class specified for each correction difference of the character string in the character string difference management table. FIG. 11 shows an example of the data structure of the character string difference management table. FIG. 11A shows a character string difference management table storing correction methods and differences corresponding to the character string correction differences DIF21 and DIF22 included in the section SEC2 of FIG. FIG. 11B shows a character string difference management table storing correction methods and differences corresponding to the character string correction differences DIF31 to DIF35 included in the section SEC3 of FIG.

  Next, the process for the correction difference of the combination information by the correction type determination unit 30 will be described.

  First, the correction type determination unit 30 specifies a correction method for each correction difference of the combination information included in the comparison result of the combination information received from the image information difference extraction unit 22. The modification difference of the combination information is classified as “addition” or “deletion” as the modification method. Specifically, the correction type determination unit 30 classifies the correction method of the combination information included only in the old version report (the combination information included in the above-described difference set P / A) as “deleted” and includes only in the new version report. Classification information (combination information included in the above-described difference set A / P) is classified as “addition”.

  Specifically, the modification difference of the combination information in FIG. 3 will be described as an example. The modification difference of the combination information (IMG2, “bronch bronchial image” inside) included only in the old version report (included in the difference set P / A) and The correction method of (IMG3, “”) is “deletion”. Further, the correction method of the combination information correction difference (IMG2, “cavity inside”) included only in the new version report (included in the difference set A / P) is “addition”.

  Also, the modification type determination unit 30 classifies each modification difference of the combination information as a “key image” as a class of differences. As described above, the correction method and the type of difference are specified for each correction difference of the combination information.

  When the correction type and the difference class are specified for each of the correction differences (the correction difference of the character string and the correction difference of the combination information), the correction type determination unit 30 determines each correction difference between the correction method and the difference class. Create a correction type determination table after classifying and tabulating based on combinations. FIG. 12 shows an example of the data structure of the correction type determination table. The above analysis is performed on the basis of the correction differences (character string correction differences and combination information correction differences) shown in the examples of FIGS. The result is entered as data.

  A method of tabulating the correction type determination table by the correction type determination unit 30 will be specifically described with reference to FIG. For example, from the data in the character string difference management table shown in FIG. 11A, there is one correction difference whose correction method is “replacement (deletion)” and whose difference type is “other words / phrases”. One correction difference that is “replacement (addition)” and whose difference type is “other words” is tabulated. Similarly, from the data in the character string difference management table shown in FIG. 11B, there are two correction differences in which the correction method is “addition” and the difference type is “findings”, and the correction method is “addition”. ”And the difference class is“ other words ”, and the correction method is“ delete ”and the difference class is“ fragment sentence ”. Note that the correction difference of the combination information is similarly tabulated based on the correction method and the type of difference.

  The correction type determination unit 30 transmits the created correction type determination table and the comparison results received from the difference extraction unit 20 (character string comparison results and combination information comparison results) to the transmission determination unit 40.

  The transmission determination unit 40 analyzes the pattern of the combination of correction differences collected in the received correction type determination table, thereby determining the type of correction (hereinafter referred to as “correction type”) applied between the old version report and the new version report. ) Can be determined (details of the analysis processing relating to the specification of the correction type will be described later).

  Further, the transmission determination unit 40 refers to the transmission determination table, and compares the comparison result (the correction made between the old version report and the new version report) received from the correction type determination unit 30 according to the specified correction type. It is determined whether or not to transmit the interpretation report to the primary interpretation doctor who created the interpretation report. The details of the process related to the specification of the correction type by the transmission determination unit 40 and the process of determining whether to transmit a comparison result (including one or more correction differences) according to the correction type will be described below.

(Identification of correction type)
The transmission determination unit 40 analyzes the types of correction differences (classification combining a correction method related to correction differences and a class of differences) collected in the correction type determination table, and specifies the correction type. Specifically, pattern data (hereinafter referred to as “pattern data”) of a correction type determination table is prepared in advance for each correction type, and the correction type determination table received from the correction type determination unit 30 is assigned to each correction type. The correction type is specified by comparing with the corresponding pattern data.

  Examples of the correction type include “correction of typographical error”, “correction of tenanoha”, “correction of finding sentence”, “correction of fragment sentence”, and the like. The contents of each correction type described above and an example of pattern data corresponding to the correction type will be described below.

The correction of an error is a correction type indicating that an error such as a part or a disease name has been corrected. FIG. 13 shows an example of pattern data in which the correction type indicates “correction of error”. In FIG. 13, only one of Xa, Ya, and Za is assumed to be 1. When the received correction type determination table satisfies all of the following <Condition C-1>, <Condition C-2>, and <Condition C-3> as in the pattern data shown in FIG. In addition, it is determined that the correction type is an incorrect correction.
<Condition C-1> The correction method includes only a correction difference of “replacement (deletion)” or “replacement (addition)”.
<Condition C-2> Regarding a correction difference whose correction method is “replacement (deletion)”, the type of difference is “erroneous”.
<Condition C-3> Regarding a modified difference whose modification method is “replacement (addition)”, the class of the difference is any one of “part”, “disease name”, and “other words”.

The correction of Tenenoh is a correction type indicating that an error relating to the usage of an auxiliary word such as a particle or an auxiliary verb has been corrected. FIG. 14 shows an example of pattern data in which the correction type indicates “correction of TENOHA”. In FIG. 14, it is assumed that Xb indicates a value of 1 or more. FIG. 14A shows a pattern in the case where unnecessary attached words are deleted. FIG. 14B shows a pattern in a case where a necessary attached word is added. FIG. 14C shows a pattern when an attached word that is misused is corrected. When the received correction type determination table satisfies <Condition D-1> shown below as in the pattern data shown in FIG. To do.
<Condition D-1> The difference class includes only the modified difference of the “attached word”.

  Note that the pattern of FIG. 14A to FIG. 14C is not limited as long as <Condition D-1> is satisfied. For example, the case where the difference class is an attached word and the correction method is an additional correction difference and the difference class is an attached word and the correction method is a deletion correction difference is also included. As described above, as long as the type of difference regarding the included correction difference is only an attached word, a plurality of types of correction differences with different correction methods may be included.

  The correction of the finding sentence is a correction type indicating that a correction to the finding sentence (a sentence including a part word, an event word, and an existence word) is included between the old version report and the new version report. The fragment sentence correction is a modification type indicating that a modification to the fragment sentence is included between the old version report and the new version report. The fragment sentence refers to a character string that includes a plurality of types of morphemes and does not include at least one of a part word, an event word, and an existence word.

  FIG. 15 shows an example of pattern data in which the correction type indicates “correction of finding sentence” or “correction of fragment sentence”. FIG. 15A shows a pattern when an unnecessary finding sentence or fragment sentence is deleted. FIG. 15B shows a pattern when an unnecessary finding sentence or fragment sentence is additionally written. FIG. 15C shows a pattern when the finding sentence or the fragment sentence is corrected. In FIG. 15, it is assumed that either Xc or Xd indicates a value of 1 or more. When Xc indicates a value of 1 or more, FIG. 14 shows an example of pattern data whose correction type is “correction of finding text”. Further, when Xd indicates a value of 1 or more, FIG. 15 shows an example of pattern data whose correction type is “correction of fragment sentence”.

The transmission determination unit 40 determines that the correction type is correction of the finding sentence when the received correction type determination table satisfies <Condition E-1> shown below as in the pattern data shown in FIG. If <Condition F-1> is satisfied, it is determined that the modification type is modification of a fragment sentence.
<Condition E-1> At least one or more correction differences whose difference type is “finding sentence” are included.
<Condition F-1> At least one or more correction differences whose difference type is “fragment sentence” are included.
If both <Condition E-1> and <Condition F-1> are satisfied, it may be determined that there are a plurality of correction types (modification of observation sentence and modification of fragment sentence). It may be determined that one of the correction types (for example, correction of a finding sentence).

  Moreover, as long as <Condition E-1> or <Condition F-1> is satisfied, the pattern is not limited to the patterns shown in FIGS. For example, if the difference class is an observation sentence or a fragment sentence, the correction method is an additional correction difference, and the difference class is an observation sentence or a fragment sentence, and the correction method is a deletion correction difference. Applicable. As described above, if the type of difference regarding the correction difference is a finding sentence or a fragment sentence, a plurality of types of correction differences with different correction methods may be included.

(Determination of transmission of comparison results)
When determining the correction type, the transmission determination unit 40 refers to the transmission determination table and determines whether or not to transmit the comparison result to the primary interpretation doctor based on the correction type. FIG. 16 shows an example of the data structure of the transmission determination table. In this case, the transmission determination unit 40 determines that the comparison result is to be transmitted when the correction type is “correction of finding sentence” or “correction of fragment sentence”, and “correction of error” or “ In the case of “correction”, it is determined that the comparison result is not transmitted.

  As a result, for example, in the case of a correction that does not include meaningful content regarding the appearance of the text, such as "correction of typographical error" or "correction of Tenenoh", it is determined that the correction is less important, and the primary It becomes possible to suppress the transmission of the comparison result to the interpretation doctor. As a result, the primary interpreting doctor can check only important corrections, and the burden on the primary interpreting doctor to check corrections by the secondary interpreting doctor can be reduced.

  It should be noted that the contents of the transmission determination table can be changed by the operator (radiologist). In FIG. 16, the comparison result is transmitted when the correction type is “correction of finding sentence” or “correction of fragment sentence”. For example, the comparison result is only when the correction type is “correction of finding sentence”. It is also possible to change the setting to transmit the comparison result or the setting to transmit the comparison result for all the correction types. As for the setting change method, for example, a user interface such as a check box may be provided in the operation unit (not shown) so that whether or not to transmit the comparison result for each correction type can be changed.

  The transmission determination unit 40 transmits the comparison result determined to be transmitted to the operator and the transmission determination table corresponding to the comparison result to the transmission content storage unit 50.

  The transmission content storage unit 50 includes a storage unit that stores the comparison result received from the transmission determination unit 40 and the transmission determination table. When the transmission content storage unit 50 receives the comparison result and the transmission determination table from the transmission determination unit 40, the version management of the information for identifying the old version report corresponding to the comparison result and the information of the interpretation doctor who created the old version report The information is extracted from the unit 12 and stored together with the comparison result and the transmission determination table received from the transmission determination unit 40.

  The correction content display method selection unit 60 has means for causing the display unit (not shown) to display the correction difference included in the comparison result stored in the transmission content storage unit 50. The correction content display method selection unit 60 receives an operation related to the display of a display target version (for example, a new version report) from the interpretation doctor, and is performed between the version (for example, the old version report) created by the interpretation doctor and the display target version. Are displayed based on the comparison result stored in the transmission content storage unit 50.

  The modification content display method selection unit 60 switches the display mode depending on whether the modification difference to be displayed is the modification difference of the character string or the modification difference of the combination information, and causes the display unit to display the difference. FIG. 17 shows an example of the display when displaying the correction difference of the character string. FIG. 18 shows an example of display in the case of displaying the modification difference of the combination information.

When displaying the correction difference of the character string as shown in FIG. 17, the correction content display method selection unit 60 sequentially reads the character string comparison result (character string comparison result in FIG. 2) for each line. The display is output by processing according to the rules.
<Rule B-1> Lines with “” (half-width space) at the beginning of the line (character strings that are not corrected between the old version report and the new version report) are output as they are.
<Rule B-2> Lines with “+” added at the beginning of the line (character strings added between the old version report and the new version report) are highlighted in bold and output.
<Rule B-3> Lines with “-” added to the beginning of the line (character strings deleted between the old version report and the new version report) are highlighted with bold letters, and are subjected to strikethrough and output.
<Rule B-4> The symbol “\ n” is output as a line feed code.

  Further, when displaying the correction difference of the combination information as shown in FIG. 18, the correction content display method selection unit 60 first includes the combination information (related to the deleted image) included only in the old version report and the new version report. Each combination information (related to the additional image) is extracted. The combination information regarding the deleted image is extracted from the difference set P / A included in the comparison result of the combination information. Further, the combination information regarding the additional image is extracted from the difference set A / P included in the comparison result of the combination information.

  The correction content display method selection unit 60 displays an image corresponding to the extracted combination information. At this time, the correction content display method selection unit 60 displays the highlighting method in a different manner depending on whether the combination information is combination information related to the additional image or combination information related to the deleted image. As a method for emphasizing and displaying an image, a method for displaying an image with a border is well known. In the example of FIG. 18, the method of bordering (for example, color) is changed and displayed for the added image and the deleted image.

  In addition, the correction content display method selection unit 60 adds a correction that needs to be transmitted to the primary interpreting doctor by the secondary interpreting doctor to the interpretation report created by the primary interpreting doctor. Alternatively, the fact that the correction has been made may be transmitted to the primary interpreting doctor. In this case, the correction content display method selection unit 60 notifies the primary interpretation doctor that the correction has been made to the interpretation report when the correction difference for the interpretation report is stored in the transmission content storage unit 50. Good.

  As described above, whether or not the interpretation report creation support system according to the present invention determines the correction type from the content of the correction applied to the interpretation report, and transmits the correction content to the primary interpreting doctor based on the correction type. Can be determined. At this time, by setting so that only the correction type including the important correction is transmitted to the primary interpretation doctor, it is possible to suppress the transmission to the primary interpretation doctor regarding the correction for the minor defect and transmit only the important correction. It becomes possible. As a result, it is possible to reduce the burden of work for the primary interpretation doctor to confirm the correction contents.

DESCRIPTION OF SYMBOLS 10 Report management part 11 Report storage part 12 Version management part 20 Difference extraction part 21 Character information difference extraction part 22 Image information difference extraction part 30 Correction type discrimination | determination part 31 Dictionary memory | storage part 40 Transmission determination part 50 Transmission content storage part 60 Correction content display Method selection part

Claims (8)

An interpretation report creation support system comprising a report management unit that identifies and stores interpretation reports of the same examination for each version,
A difference extraction unit that extracts different descriptions generated as a result of correction for different interpretation reports of the same examination version,
A correction method is specified based on the different description aspect of the correction difference, and the correction difference includes a plurality of observation sentences by analyzing the meaning content of the character string of the correction difference and the aspect of the sentence structure. A correction content extraction unit for identifying which of the types corresponds,
Based on the combination of the specified correction method and the type, with reference to the table showing the classification of the correction type based on the combination of the correction method and the type of correction difference, the correction type including correction of the finding sentence An interpretation report creation support system comprising: a transmission determination unit that specifies any one of them and determines whether or not to transmit the correction difference according to the specified correction type. The difference extraction unit compares the description of the first version with the description of the second version, and extracts the description that differs between the first version and the second version as one or more correction differences. And
The correction content extraction unit identifies a correction method for each of the correction differences by detecting which of the first version and the second version is included. Reading interpretation report creation support system.   The difference extraction unit compares the character string described in the first version with the character string described in the second version, and the character string included only in the first version as the correction difference The image interpretation report creation support system according to claim 2, further comprising a character information difference extraction unit that extracts either or both of the correction difference of the character string and the correction difference of the character string included only in the second version. .   The correction content extraction unit further includes a first dictionary storage unit that stores a dictionary including a character string described in a report and a classification based on the semantic content of the character string, and the correction difference of the character string is calculated as the first difference. 4. The interpretation report generation according to claim 1, wherein the meaning content of the corrected difference character string is analyzed by comparing with a character string stored in one dictionary storage unit. Support system.   The correction content extraction unit decomposes the correction difference of the character string into morphemes, and then analyzes an aspect of the sentence structure based on a combination of morphemes included in the correction difference of the character string. The interpretation report creation support system according to any one of claims 1 to 3.   The correction content extraction unit further includes a second dictionary storage unit that stores a dictionary including a character string in morpheme units described in a report and a classification based on the semantic content of the character string in morpheme units, and the analysis The classification based on the semantic content corresponding to each of the decomposed morphemes is performed by comparing the decomposed morphemes with character strings in morpheme units stored in the second dictionary storage unit. The type of correction difference including a sentence sentence that constitutes a part of a finding sentence and a finding sentence based on a combination of classifications based on the semantic content corresponding to each of the morphemes included in the modification difference of the character string The interpretation report creation support system according to claim 5, wherein the interpretation report creation support system is identified and classified into any one of the above. The difference extraction unit, the combination information and the image associated with the character string described in the first plate and the second plate each of the interpretation report and the character string are combined, and extracted as a list, said combination by comparing the list of previous list SL said combination information extracted from the first plate and the combination information extracted from the second plate, which as said modified differential is included only in said first plate An image information difference extraction unit that extracts either or both of a correction difference of information and a correction difference of the combination information included only in the second version;
The interpretation report creation support system according to claim 1, wherein the correction content extraction unit classifies the correction difference of the combination information for each predetermined type of the correction difference. A display for displaying the correction difference, and a correction content display for changing the display form according to the classification for each type of the correction difference and displaying the correction difference on the display unit based on the determination result of the transmission determination unit The interpretation report creation support system according to claim 1, further comprising a method selection unit.