JP7473314B2 - Medical information management device and method for adding metadata to medical reports - Google Patents

Description

The present invention relates to a medical information management device and a method for adding metadata to medical reports that can add metadata to medical reports, such as radiology reports and pathology reports, in order to quickly and accurately search for medical image information associated with the medical reports.

Traditionally, in the medical field, technology has become widespread to convert many radiology reports into electronic data and store them in database systems such as PACS (Picture Archiving and Communication System). In this database system, the electronic radiology reports are stored in association with the image data that was the subject of the radiology interpretation.

An image interpretation report is a document in which the doctor interpreting the images, based on his or her medical and clinical knowledge, thinks about what is happening based on information about the patient and the medical images that were taken. It also describes what abnormalities the doctor found in the medical images and what illness the doctor thinks the patient may have.

For this reason, in order to search for medical images, there are methods that generate structured data by adding metadata to the text data written in the findings section of the radiology report (see Patent Documents 1 to 3).

JP 2008-052544 A JP 2011-028730 A JP 2016-151827 A

However, the style of the findings section of many radiology reports varies depending on the disease in question and the physician who wrote it. In addition, because radiology reports are assumed to be read by radiology physicians who have knowledge of medicine and clinical practice, they may use specialized terminology that assumes knowledge of medicine and clinical practice, or omit certain expressions.

In addition, although the findings section of the radiology report is written in natural language, the writing style varies depending on the radiologist, the rules of grammar and other aspects are vague, and the order in which expressions appear and the order in which information and sentences are written may be expressed in a different order.

For this reason, it is conceivable that radiologists could generate structured data by tagging the text data of findings, but this would place a heavy burden on the radiologist and would take a long time to generate the structured data.

As described in Patent Documents 1 to 3, it is possible to automatically generate structured data from text data of findings. However, if the automatically added metadata is influenced by the ambiguity of the text data of findings and does not have appropriate attribute values, there is a problem that searching for medical images takes a long time and the search accuracy deteriorates.

In particular, in recent years, a huge number of medical reports have come to be stored on cloud servers, creating a demand for speed and high accuracy when searching medical images.

The present invention has been made to solve the problems of the conventional technology described above, and aims to provide a medical information management device and a method for adding metadata to medical reports that can add metadata to medical reports, such as radiology reports and pathology diagnosis reports, in a fast and accurate manner.

In order to solve the above problems, the present invention provides a medical information management device that is associated with medical image information and that adds search metadata to a medical report that describes text data of a finding, and is characterized in comprising: a medical dictionary that stores medical terminology including a standard description for search; an element description extraction unit that analyzes the text data of the finding based on the medical dictionary to extract an element description that matches a plurality of attribute items for search from the text data of the finding and extracts an element description that includes a part of the plurality of attribute items for search as an ambiguous element description ; and a structured data generation unit that converts the element description into a standard description for search based on the medical dictionary , generates structured data in which the converted standard description for search is added to the text data of the finding as metadata associated with the element description, and, when the ambiguous element description is extracted, and when a description combined with another element description or another ambiguous element description becomes the standard description for search, generates structured data in which the standard description for search is added to the text data of the finding as metadata associated with the ambiguous element description .

The present invention is also characterized in that, in the above invention, the structured data generation unit generates structured data in which the search standard description is added to the text data of the findings as coded metadata.

The present invention is also characterized in that, in the above invention, it includes an image search unit that searches for medical image information associated with the medical report based on metadata in the medical report.

In the above invention, the image search unit searches for the medical image information using metadata in the medical report and metadata of the medical image information.

The present invention also provides a method for adding metadata to a medical report, which is associated with medical image information and includes text data of a finding, and which adds metadata for search to a medical report, the method comprising: a medical dictionary storing medical terminology including a standard description for search; an element description extraction step for analyzing the text data of the finding based on the medical dictionary to extract an element description that matches a plurality of attribute items for search from the text data of the finding , and extracting an element description including a part of the plurality of attribute items for search as an ambiguous element description; and a structured data generation step for converting the element description into a standard description for search based on the medical dictionary, generating structured data in which the converted standard description for search is added to the text data of the finding as metadata associated with the element description, and, when the ambiguous element description is extracted, a description in combination with another element description or another ambiguous element description becomes the standard description for search, generating structured data in which the standard description for search is added to the text data of the finding as metadata associated with the ambiguous element description .

In the above invention, the structured data generation step generates structured data in which the search standard description is added to the text data of the findings as coded metadata.

The present invention makes it possible to add metadata to medical images associated with medical reports, such as radiology reports and pathology reports, in a fast and accurate manner.

FIG. 1 is a functional block diagram showing the configuration of a medical information management system including a medical information management server. FIG. 2 is a diagram showing an example of a template screen of an image interpretation report. FIG. 3 is a diagram showing an example of search attribute items. FIG. 4 is a diagram showing an example of text data and an example of extracted element descriptions. FIG. 5 is a flowchart showing a procedure for adding metadata to an image interpretation report by the control unit.

Below, a preferred embodiment of the medical information management device and method for adding metadata to a medical report according to the present invention will be described in detail with reference to the attached drawings.

<Overview of medical image management system>
First, a configuration of a medical information management system 1 including a medical information management server 10 corresponding to a medical information management device according to the present embodiment will be described. FIG. 1 is a functional block diagram showing the configuration of the medical information management system 1 including the medical information management server 10.

As shown in FIG. 1, in the medical information management system 1, a medical information management server 10, a local server 11, and a terminal device 15 are connected to a network N. A modality 12 and an image reading device 13 are also connected to the local server 11. The medical information management server 10 is, for example, a server formed on the cloud. The local server 11, the modality 12, and the image reading device 13 are, for example, located in a medical facility such as a hospital. The terminal device 15 is, for example, located in a medical facility or a research institute.

The modality 12 is an image generating device such as CR (Computed Radiography), MRI (Magnetic Resonance Imaging), or CT (Computed Tomography). The modality 12 transmits attribute information of the examined patient and the like along with the generated medical image to the local server 11. The local server 11 adds the attribute information to the medical image as metadata.

<Image Reading Device>
The image reading device 13 is a device for reading medical image information, in which metadata is added to medical images, from the local server 11 and allowing a doctor to read the images. The image reading device 13 has an image reading report creation support unit 14. The image reading report creation support unit 14 displays an image reading report template and supports the creation of an image reading report.

2 is a diagram showing an example of a template screen of the image reading report 41 to be created. As shown in FIG. 2, the template screen displays, for example, "Report ID", "Patient name", "Examination date", "Image reading doctor", "Image reading date", "Findings column", and "Image 1", "Image 2", and "Image 3" which are links to medical images. The image reading report creation support unit 14 automatically assigns the "Report ID", "Patient name", "Examination date", "Image reading date", and the links "Image 1", "Image 2", and "Image 3" based on the loaded medical image information. When logging in to the image reading device 13, the "image reading doctor" may automatically assign the ID of the image reading doctor by reading the ID of the image reading doctor in advance. The image reading doctor freely writes the image reading results in the natural language text data D1 in the findings column E1 of the template screen while displaying the medical image of the loaded medical image information in a separate window. The local server 11 sends the created image reading report 41 to the medical information management server 10 via the network N, and causes the image reading report 41 to be stored in the medical information management server 10. When a large amount of medical image information is sent from the modality 12, the local server 11 may transfer the medical image information to an image reading center (not shown), and the image reading device in the image reading center may create the image reading report 41. In this case, the image reading device in the image reading center sends the created image reading report 41 to the medical information management server 10 via the network N.

<Medical information management server>
The medical information management server 10 has an input/output unit 20, a communication unit 30, a storage unit 40, and a control unit 50.

The input/output unit 20 is an input/output interface such as a touch panel display that performs various operation inputs and display outputs. The communication unit 30 is a communication interface for communicating with the local server 11 and the terminal device 15 via the network N.

The storage unit 40 is a storage device consisting of a hard disk drive, non-volatile memory, etc., and stores image interpretation reports 41, medical image information 42, and a medical dictionary 43. The medical dictionary 43 is a thesaurus dictionary that stores medical terminology including a search standard description 44 and includes relationships such as synonyms. The medical dictionary 43 may be a general dictionary, but in this embodiment, it is a unique dictionary that includes past terms written in the findings column. The search standard description 44 is, for example, a term used in ICD10 (International Classification of Diseases). In this case, a code 45 is associated with the search standard description 44.

The control unit 50 controls the entire medical information management server 10, and has a reception unit 51, an element description extraction unit 52, a structured data generation unit 53, and an image search unit 54. The control unit 50 stores programs corresponding to these functional units in a storage device such as a non-volatile memory or a magnetic disk device, and loads these programs into memory and executes them on the CPU to execute the corresponding processes.

The reception unit 51 receives various information such as the image interpretation report 41 from the local server 11 and a search request from the terminal device 15 via the communication unit 30.

When the element description extraction unit 52 receives the image interpretation report 41, it analyzes the findings text data D1 based on the medical dictionary 43 and extracts multiple element descriptions related to multiple search attribute items from the findings text data D1.

As shown in FIG. 3, the search attribute items are, for example, the part, the findings, the disease name, the score, etc. The attribute value of the part is, for example, the liver area "S7". The attribute value of the findings is, for example, ground glass opacity, etc. The attribute value of the disease name is, for example, the disease name in ICD10. The attribute value of the score is, for example, positive, ambiguous, negative, etc.

The structured data generating unit 53 converts the extracted element description into a standard description for search, and generates structured data in which the converted standard description for search is added to the text data D1 of the findings as metadata associated with the element description. The structured data generating unit 53 converts the text data D1 into an HTML document in which attribute items and attribute values for search are added, for example.

When the extracted element description is a standard description for search, the structured data generation unit 53 adds the standard description for search to the element description as metadata. Also, when the extracted element description is not a standard description for search but there is a standard description for search within the spelling variation range, the structured data generation unit 53 adds the standard description for search within the spelling variation range as metadata to the element description. For example, when the element description is "UAP" or "uAP", there is a standard description for search that is "unstable angina" within the spelling variation range, so metadata indicating "unstable angina" is added to the element description "UAP" or "uAP".

Here, if the extracted element description is a description that includes part of the standard description for search, the element description extraction unit 52 extracts it as an ambiguous element description. Then, if an ambiguous element description is extracted and a description combined with another element description or another ambiguous element description becomes a standard description for search, the structured data generation unit 53 adds this standard description for search as metadata associated with the ambiguous element description.

Figure 4 is a diagram showing an example of text data D1 and an example of extracted element descriptions. As shown in Figure 4, the element description extraction unit 52 extracts liver, cyst, possible, cancer, possible from the text data D1 "Possible liver cyst, possible cancer". The element description "liver" is an attribute item for searching a part and is also a standard description for searching, so it is added as it is to "liver" as metadata. In addition, since "cyst" is neither a disease name nor a part, it is extracted as an ambiguous element description. Then, "cyst" becomes the standard description for searching a disease name "liver cyst" when combined with "liver", a part of another element description "liver", so the disease name "liver cyst" is added as metadata to the ambiguous element description "cyst". Note that the element description "cancer" is also extracted as an ambiguous element description, but since there is no standard description for searching in combination with other element descriptions, no metadata is added. In addition, "ambiguous" is added as metadata to the element description "possible" as an attribute value for the score.

The above-mentioned standard search description may be coded using code 45. For example, if the disease name is "liver cyst", the ICD10 code "K768" is added as metadata. In the medical field, code 45 is often used to identify the standard search description, so adding code 45 as metadata allows for faster and more accurate searches.

When the image search unit 54 receives a search request with search conditions added from the terminal device 15, it searches for medical images in the medical image information 42 associated with the image interpretation report 41 based on the metadata in the image interpretation report 41, and presents the search results to the terminal device 15. Note that when searching for medical images, the image search unit 54 may search using the metadata of the medical image information 42 in addition to the metadata in the image interpretation report.

<Terminal Device>
The terminal device 15 searches for medical images stored in the medical information management server 10 using metadata added to the image interpretation report 41. The terminal device 15 has an image search support unit 16. The image search support unit 16 displays a search template for inputting attribute values for search attribute items, and supports the search for medical images.

When the above-mentioned medical information management server 10 converts the text data D1 of the findings in the radiology report 41 into structured data with added metadata, it takes into consideration that medical images will be searched for and uses standard search descriptions as attribute values of the metadata, making it possible to search for medical images quickly and accurately.

<Adding Metadata to Radiography Reports>
Fig. 5 is a flowchart showing a procedure for adding metadata to an image interpretation report by the control unit 50. As shown in Fig. 5, when the reception unit 51 of the control unit 50 receives the image interpretation report 41 (step S101), the element description extraction unit 52 analyzes the text data D1 of the findings column E1 based on the medical dictionary 43 and extracts a plurality of element descriptions related to a plurality of search attribute items from the text data D1 (step S102).

Then, the structured data generation unit 53 judges whether the extracted element description is a standard description for search (step S103). If the element description is a standard description for search (step S103; Yes), the standard description for search is added to this element description as metadata (step S104), and the process proceeds to step S107.

If the element description is not a standard description for search (step S103; No), it is further determined whether or not there is a standard description for search within the spelling variation range of the extracted element description (step S105). If there is a standard description for search within the spelling variation range of the extracted element description (step S105; Yes), this standard description for search is added to the element description as metadata (step S104), and the process proceeds to step S107.

If there is no standard description for search within the range of notation variation of the extracted element description (step S105; No), it is further determined whether a description combining part of the element description with another element description or another ambiguous element description constitutes a standard description for search (step S106). If the combined description constitutes a standard description for search (step S106; Yes), this standard description for search is added to the element description as metadata (step S104), and the process proceeds to step S107.

On the other hand, if the combined description does not become a standard description for search (step S106; No), it is further determined whether there are any unprocessed element descriptions (step S107). If there are any unprocessed element descriptions (step S107; Yes), the process proceeds to step S103 and metadata is added to the next element description. On the other hand, if there are no unprocessed element descriptions (step S107; No), this process ends.

In the above embodiment, the addition of metadata to the radiology report 41 has been described, but the present invention is not limited to this and can also be applied to pathology diagnosis reports. In other words, the present invention is applied to medical reports associated with medical images.

In addition, in the above embodiment, it is assumed that the element description extraction unit 52 and the structured data generation unit 53 perform processing based on a rule base, but this is not limited to the above. Metadata may be added to the radiology report using deep learning.

Note that the configurations illustrated in the above examples are merely functional schematics, and do not necessarily have to be physically configured as illustrated. In other words, the form of distribution and integration of each device is not limited to that illustrated, and all or part of the devices can be functionally or physically distributed and integrated in any unit depending on various loads, usage conditions, etc.

The medical information management device and method for adding metadata to medical reports according to the present invention are useful for adding metadata to quickly and accurately search for medical image information associated with medical reports such as radiology reports and pathology diagnosis reports.

REFERENCE SIGNS LIST 1 Medical information management system 10 Medical information management server 11 Local server 12 Modality 13 Image reading device 14 Image reading report creation support unit 15 Terminal device 16 Image search support unit 20 Input/output unit 30 Communication unit 40 Memory unit 41 Image reading report 42 Medical image information 43 Medical dictionary 44 Search standard description 45 Code 50 Control unit 51 Reception unit 52 Element description extraction unit 53 Structured data generation unit 54 Image search unit D1 Text data E1 Findings column N Network

Claims (6)

A medical information management device that adds search metadata to a medical report that is associated with medical image information and contains text data of findings, comprising:
a medical dictionary containing medical terminology including standard descriptions for searching;
an element description extraction unit that analyzes the text data of the findings based on the medical dictionary, extracts element descriptions that match a plurality of search attribute items from the text data of the findings , and extracts element descriptions that include a part of the plurality of search attribute items as ambiguous element descriptions ;
and a structured data generation unit that converts the element description into a standard description for search based on the medical dictionary , generates structured data in which the converted standard description for search is added to the text data of the finding as metadata associated with the element description, and when the ambiguous element description is extracted, if another element description or a description combined with another ambiguous element description becomes the standard description for search, generates structured data in which the standard description for search is added to the text data of the finding as metadata associated with the ambiguous element description . 2. The medical information management device according to claim 1, wherein the structured data generating unit generates structured data in which the search standard description is added to the text data of the findings as coded metadata. 3. The medical information management apparatus according to claim 1, further comprising an image search unit for searching for medical image information associated with the medical report based on metadata in the medical report. 4. The medical information management apparatus according to claim 3 , wherein the image search unit searches for the medical image information by using metadata in the medical report and metadata of the medical image information. A metadata assignment method executed by a medical information management device , which assigns search metadata to a medical report associated with medical image information and including text data of findings, comprising:
The medical information management device includes a medical dictionary storing medical terminology including a standard description for search;
an element description extraction step in which the medical information management device analyzes the text data of the findings based on the medical dictionary, extracts element descriptions that match a plurality of search attribute items from the text data of the findings , and extracts element descriptions that include a part of the plurality of search attribute items as ambiguous element descriptions ;
and a structured data generating step in which the medical information management device converts the element description into a standard description for search based on the medical dictionary , and generates structured data in which the converted standard description for search is added to the text data of the finding as metadata corresponding to the element description , and when the ambiguous element description is extracted, if another element description or a description combined with another ambiguous element description becomes the standard description for search, generates structured data in which the standard description for search is added to the text data of the finding as metadata corresponding to the ambiguous element description . 6. The metadata adding method according to claim 5, wherein said structured data generating step generates structured data in which said search standard description is added to said text data of said findings as coded metadata.

Images