Medical examination report interpretation method, device and electronic device

Abstract

The present invention provides a method, device and electronic device for interpreting medical examination reports, which obtain unstructured medical examination reports; based on a preset medical case model, the unstructured medical examination reports are structured by using a natural language processing method; the structured data in the structured medical examination reports are matched with the standard data in the pre-acquired medical knowledge database to obtain a matching result; based on the matching result, an interpretation file corresponding to the unstructured medical examination report is generated. This method uses a medical case model and natural language processing technology to structure the unstructured medical examination reports and match them with the standard data in the medical knowledge database, and generates an interpretation file based on the matching results, which can meet the patient's demand for automatic interpretation of unstructured text-based examination reports.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a method, device and electronic equipment for interpreting medical examination reports. Background Art

[0002] During the diagnosis and treatment process, medical imaging can be used to check the patient's internal body conditions. For most patients, due to the lack of professional medical knowledge, when they receive the medical imaging examination report, they often do not know what the contents described in the report mean. Although there is a lot of knowledge that can be searched and referenced on the Internet, due to the different individual conditions of each patient, it is often not applicable to everyone, and even some misleading information can have a negative impact. In related technologies, the examination report can be interpreted through OCR (Optical Character Recognition) recognition technology, but this method can only interpret structured laboratory examination reports. For unstructured text-based medical imaging examination reports, pathology examination reports, etc., automatic interpretation cannot be achieved, resulting in the inability to meet patients' needs for automatic interpretation of unstructured text-based examination reports. Summary of the invention

[0003] The purpose of the present invention is to provide a method, device and electronic device for interpreting medical examination reports to meet the patient's demand for automatic interpretation of examination reports in unstructured text form.

[0004] The present invention provides a method for interpreting a medical examination report, the method comprising: obtaining an unstructured medical examination report; based on a preset medical case model, using a natural language processing method to structure the unstructured medical examination report; matching the structured data in the structured medical examination report with the standard data in a pre-acquired medical knowledge database to obtain a matching result; and generating an interpretation file corresponding to the unstructured medical examination report based on the matching result.

[0005] Furthermore, based on the preset medical case model, the step of structuring the medical examination report by using natural language processing includes: using natural language processing to store the unstructured data in the unstructured medical examination report according to the structure predefined by the medical case model to obtain a structured medical examination report.

[0006] Furthermore, the structured data in the structured medical examination report after structured processing is matched with the standard data in the medical knowledge database acquired in advance, and the steps of obtaining the matching result include: screening out the indicator characteristic values ​​in digital form from the structured data in the structured medical examination report after structured processing; matching the structured fields associated with the indicator characteristic values ​​in digital form in the structured medical examination report with multiple fields in the medical knowledge database acquired in advance to obtain standard fields with matching relationships; standardizing the indicator names associated with the indicator characteristic values ​​in digital form to obtain standardized indicator names; and determining the matching result based on the indicator characteristic values ​​in digital form, the standard fields with matching relationships and the standardized indicator names.

[0007] Furthermore, based on the indicator characteristic values ​​in digital form, standard fields with matching relationships and standardized indicator names, the steps of determining matching results include: obtaining standard characteristic values ​​from a medical knowledge database based on standard fields with matching relationships and standardized indicator names; comparing the indicator characteristic values ​​in digital form with the standard characteristic values ​​to determine abnormal indicator characteristic values; and determining matching results based on the abnormal indicator characteristic values.

[0008] Furthermore, based on the abnormal indicator characteristic value, the step of determining the matching result includes: determining a first inspection result corresponding to the abnormal indicator characteristic value based on the abnormal indicator characteristic value; filtering out a specified characteristic value containing specified characters from other indicator characteristic values ​​except the indicator characteristic values ​​in digital form in the structured data of the medical examination report after structured processing; obtaining a specified indicator name and an inspection part name corresponding to the specified characteristic value; based on the specified characteristic value, spelling the specified indicator name and the inspection part name according to preset spelling rules to obtain a spelling result; based on the spelling result, determining a second inspection result corresponding to the specified characteristic value; and determining a matching result based on the first inspection result and the second inspection result.

[0009] Furthermore, based on the first examination result and the second examination result, the step of determining the matching result includes: obtaining specified information in an unstructured medical examination report; wherein the specified information includes at least one of the following: gender information, age information; based on the specified information, the first examination result and the second examination result, determining the matching result.

[0010] Furthermore, based on the specified information, the first inspection result and the second inspection result, the step of determining the matching result includes: screening out abnormal inspection results that meet the specified information from the first inspection result and the second inspection result; obtaining interpretation information that matches the abnormal inspection result; wherein the interpretation information includes at least one of the following: medical explanations and health guidance suggestions corresponding to the abnormal inspection result; and determining the matching result based on the abnormal inspection result and the interpretation information.

[0011] The present invention provides a device for interpreting medical examination reports, which includes: an acquisition module for acquiring an unstructured medical examination report; a processing module for structuring the unstructured medical examination report based on a preset medical case model and using a natural language processing method; a matching module for matching the structured data in the medical examination report after the structured processing with the standard data in a medical knowledge database acquired in advance to obtain a matching result; and a generation module for generating an interpretation file corresponding to the unstructured medical examination report based on the matching result.

[0012] An electronic device provided by the present invention includes a processor and a memory, wherein the memory stores machine executable instructions that can be executed by the processor, and the processor executes the machine executable instructions to implement any of the above-mentioned medical examination report interpretation methods.

[0013] The present invention provides a machine-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement any of the above-mentioned medical examination report interpretation methods.

[0014] The medical examination report interpretation method, device and electronic device provided by the present invention obtain unstructured medical examination reports; based on a preset medical case model, the unstructured medical examination report is structured by using a natural language processing method; the structured data in the structured medical examination report is matched with the standard data in the pre-acquired medical knowledge database to obtain a matching result; based on the matching result, an interpretation file corresponding to the unstructured medical examination report is generated. This method uses a medical case model and natural language processing technology to structure the unstructured medical examination report and match it with the standard data in the medical knowledge database, and generates an interpretation file based on the matching result, which can meet the patient's automatic interpretation needs for unstructured text-based examination reports. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 A flowchart of a method for interpreting a medical examination report provided by an embodiment of the present invention;

[0017] Figure 2A flowchart of another method for interpreting a medical examination report provided by an embodiment of the present invention;

[0018] Figure 3 A flowchart of another method for interpreting a medical examination report provided by an embodiment of the present invention;

[0019] Figure 4 A flowchart of another method for interpreting a medical examination report provided by an embodiment of the present invention;

[0020] Figure 5 A schematic diagram of the structure of a medical examination report interpretation device provided by an embodiment of the present invention;

[0021] Figure 6 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] Medical imaging is the main method used to view the internal condition of the patient's body during the diagnosis and treatment process, and even has clinical significance as the gold standard for diagnosis of certain diseases. In the related technology, the data of structured laboratory test reports can be identified and classified based on OCR technology, and abnormal indicator information can be judged and extracted through keyword indicators, and the laboratory test report interpretation report can be generated in combination with the knowledge base; the steps of this method generally include: OCR recognition of laboratory test reports in image format; capturing abnormal indicator information and generating electronic test reports; combining with expert knowledge base to automatically generate recommended information; generating laboratory test report interpretation files based on recommended information. However, the biggest problem with this method is that it can only interpret structured laboratory test reports, and does not achieve the effect of automatic interpretation for unstructured text-based medical imaging and pathology reports. From the application level of report interpretation, the related technology does not solve the patient's demand for automatic interpretation of unstructured text-based medical imaging and pathology reports.

[0024] Based on this, an embodiment of the present invention provides a method, device and electronic device for interpreting a medical examination report. The technology can be applied to applications that require interpretation of unstructured medical examination reports.

[0025] To facilitate understanding of this embodiment, a method for interpreting a medical examination report disclosed in an embodiment of the present invention is first described in detail; Figure 1 As shown, the method comprises the following steps:

[0026] Step S102: Obtain an unstructured medical examination report.

[0027] The above-mentioned unstructured medical examinations may include medical imaging examination reports, pathological examination reports, etc. The report content in unstructured medical examination reports is usually unstructured text data; medical examination reports are usually written according to the medical record writing standards and in accordance with the differences of each individual patient, and there are various writing styles; for example, taking medical imaging examination reports as an example, based on the basic principles of writing medical imaging examination reports, the lesion description in the examination report in the medical record usually needs to include the examined part and the nature characteristics of the examined part, where the nature characteristics may include the distribution, number, size, thickness, etc. of the lesions. In actual implementation, when it is necessary to automatically interpret the unstructured medical examination report, it is usually necessary to obtain the unstructured medical examination report first.

[0028] Step S104, based on the preset medical case model, the unstructured medical examination report is structured using natural language processing.

[0029] The above-mentioned medical case model can be understood as a structured data structure defined based on the electronic medical record data and the applicable scenario; the above-mentioned natural language processing is an important direction in the fields of computer science and artificial intelligence. It studies various theories and methods that can achieve effective communication between people and computers using natural language. Natural language processing is a science that integrates linguistics, computer science, and mathematics, that is, to let computers understand the value of the language we use in daily life. In actual implementation, in order to effectively interpret unstructured medical examination reports, natural language processing technology and medical case models can be used to structure unstructured medical examination reports to obtain structured examined parts, as well as properties such as size and thickness.

[0030] Step S106, matching the structured data in the structured medical examination report with the standard data in the pre-acquired medical knowledge database to obtain a matching result.

[0031] The above-mentioned medical knowledge database is based on evidence-based medicine and is relatively standardized knowledge formed by integrating design and associating knowledge points. In actual implementation, when the unstructured medical examination report is structured, a structured medical examination report can be obtained. The report content in the structured medical examination report is structured data, and the structured data can be matched with the standard data in the medical knowledge database, that is, an abnormal comparison is performed under unified conditions to obtain a matching result.

[0032] Step S108: Generate an interpretation file corresponding to the unstructured medical examination report based on the matching result.

[0033] The above-mentioned medical examination report interpretation method obtains an unstructured medical examination report; based on a preset medical case model, the unstructured medical examination report is structured by using a natural language processing method; the structured data in the structured medical examination report is matched with the standard data in the pre-acquired medical knowledge database to obtain a matching result; based on the matching result, an interpretation file corresponding to the unstructured medical examination report is generated. This method uses a medical case model and natural language processing technology to structure the unstructured medical examination report and match it with the standard data in the medical knowledge database, and generates an interpretation file based on the matching results, which can meet the patient's automatic interpretation needs for unstructured text-based examination reports.

[0034] The embodiment of the present invention also provides another method for interpreting medical examination reports, which is implemented on the basis of the method of the above embodiment; the method focuses on describing the specific process of structurally processing the medical examination report by using natural language processing based on a preset medical case model, such as Figure 2 As shown, the method comprises the following steps:

[0035] Step S202: Obtain an unstructured medical examination report.

[0036] Step S204, using a natural language processing method, the unstructured data in the unstructured medical examination report is stored according to a structure predefined by the medical case model to obtain a structured medical examination report.

[0037] In actual implementation, natural language processing technology can be used to store the unstructured data in the unstructured medical examination report, that is, the electronic medical record data, according to the structure pre-defined by the medical case model to obtain a structured medical examination report. For example, see a table structure shown in Table 1, which includes categories such as model, attribute, data type, etc. Table 1 takes cervical ultrasound examination as an example. The corresponding attributes may include side name, part name, presence or absence of nodules, presence or absence of multiple lesions, etc., which can be filled in with reference to the table structure to obtain a structured medical examination report.

[0038] Table 1

[0039]

[0040]

[0041] For example, the content of a medical examination report is: "The left lobe of the thyroid gland is 3.5 cm thick, and a low echo can be seen in the left lobe of the thyroid gland. The echo size is about 0.9×0.6 cm, the boundary is not clear, the shape is irregular, and a dot-like strong echo can be seen inside it, CDI, and a small amount of blood flow signal can be seen inside it. The right lobe of the thyroid gland is 3.5 cm thick, the isthmus is 0.3 cm thick, and another echo-free area can be seen near the isthmus of the right lobe. The echo size is about 0.4×0.2 cm, the boundary is clear, CDI, the thyroid parenchyma has uniform echo, the blood flow distribution in the thyroid gland is normal, and there are obvious enlarged lymph nodes in both necks." After structured processing, the structured processing results can be obtained as follows:

[0042]

[0043]

[0044]

[0045] Step S206, matching the structured data in the structured medical examination report with the standard data in the pre-acquired medical knowledge database to obtain a matching result.

[0046] Step S208: Generate an interpretation file corresponding to the unstructured medical examination report based on the matching result.

[0047] The above-mentioned method for interpreting medical examination reports obtains unstructured medical examination reports. Natural language processing is used to store the unstructured data in the unstructured medical examination report according to the structure predefined by the medical case model to obtain a structured medical examination report. The structured data in the structured medical examination report is matched with the standard data in the pre-acquired medical knowledge database to obtain a matching result. Based on the matching result, an interpretation file corresponding to the unstructured medical examination report is generated. This method uses medical case models and natural language processing technology to structure the unstructured medical examination report and match it with the standard data in the medical knowledge database. The interpretation file is generated according to the matching results, which can meet the patient's automatic interpretation needs for unstructured text-based examination reports.

[0048] The embodiment of the present invention also provides another method for interpreting medical examination reports, which is implemented on the basis of the method of the above embodiment; the method focuses on the specific process of matching the structured data in the medical examination report after structured processing with the standard data in the medical knowledge database acquired in advance to obtain the matching result, such as Figure 3 As shown, the method comprises the following steps:

[0049] Step S302: Obtain an unstructured medical examination report.

[0050] Step S304, based on the preset medical case model, the unstructured medical examination report is structured using a natural language processing method.

[0051] Step S306, filtering out the indicator characteristic values ​​in digital form from the structured data of the medical examination report after structured processing.

[0052] The above-mentioned indicator characteristic values ​​can also be understood as the property characteristic values ​​in the medical examination report, etc.; in the structured data of the medical examination report after structured processing, there may be indicator characteristic values ​​in digital form, that is, situations that can be directly compared. By establishing a matching relationship between the medical record structured field and the knowledge field, the structured property characteristics of the examination report can be compared with the normal range of the knowledge data, such as the structured property characteristics of the examination report are: the thickness of the thyroid lobe, the property characteristic value: 3.5mm, etc.; there may also be indicator characteristic values ​​in non-digital form, that is, situations that cannot be directly compared, that is, situations where the structured property characteristics of the examination report cannot be compared with the normal range of the knowledge data, such as the structured property characteristics of the examination report are: location, whether the lymph nodes are enlarged, the property characteristic value: location = neck, whether the lymph nodes are enlarged = yes, for the property characteristic values ​​that cannot be directly compared, after filtering the medical entity synonym library and standardization, they are compared to determine whether the property characteristics are normal or abnormal.

[0053] For the indicator characteristic values ​​in digital form, since they can be directly compared with the standard data in the medical knowledge database, the indicator characteristic values ​​in digital form can be screened out from the structured data of the medical examination report after structured processing.

[0054] Step S308, matching the structured fields associated with the numerical indicator characteristic values ​​in the structured medical examination report with the multiple fields in the pre-acquired medical knowledge database to obtain standard fields with matching relationships.

[0055] In actual implementation, it is necessary to match the diverse medical record data with relatively standard knowledge fields so that abnormal comparison can be performed under unified conditions. The following are examples of medical record data and knowledge data:

[0056] Medical records-examination reports:

[0057] "The left lobe of the thyroid gland is 3.5 cm thick. A low echo can be seen in the left lobe of the thyroid gland. The echo size is about 0.9×0.6 cm, the boundary is not clear, and the shape is irregular. Point-like strong echoes can be seen inside it, CDI, and a small amount of blood flow signals can be seen inside it. The right lobe of the thyroid gland is 3.5 cm thick, and the isthmus is 0.3 cm thick. Another echoless area can be seen near the isthmus of the right lobe. The echo size is about 0.4×0.2 cm, with clear boundaries, CDI. The thyroid parenchyma has uniform echoes, the blood flow distribution in the thyroid gland is normal, and there are obviously enlarged lymph nodes in both necks."

[0058] Knowledge data:

[0059] "Lie on your back with your neck tilted back:

[0060] The normal thyroid lateral lobe has an anteroposterior diameter of 1-2 cm, a left-right diameter of 2-2.5 cm, a superior-inferior diameter of 3.5-5 cm, and an anteroposterior diameter of the isthmus of 0.2-0.4 cm. It is small, densely distributed, and evenly distributed with medium echoes, and the surrounding muscle tissue is low echoes. In an abnormal state, it may be reduced, increased, or unevenly echoed.

[0061] For the cases that can be directly compared, for example, the structured medical record data is:

[0062]

[0063] The knowledge data in the medical knowledge database are:

[0064]

[0065]

[0066] By comparing the field values ​​of the structured medical record data and the knowledge data, we can establish a matching relationship between the medical record field and the knowledge field based on the fields referring to the same content by using methods such as keyword matching, as shown in Table 2 below:

[0067] Table 2

[0068]

[0069] Step S310, standardize the indicator name associated with the indicator characteristic value in digital form to obtain a standardized indicator name.

[0070] The above indicator names can be understood as the names of the property characteristics in the medical examination report. For example, continuing with the example in the previous step, the indicator name in the structured medical record data is thyroid lobe; in actual implementation, the property characteristic values ​​can be standardized based on the medical entity synonym library, for example, an indicator name before and after standardization as shown in Table 3.

[0071] Table 3

[0072] Before standardization After standardization Thyroid lobe Lateral lobe of thyroid gland Thyroid lobe Lateral lobe of thyroid gland thickness Front-back diameter thick Front-back diameter

[0073] Step S312, determining a matching result based on the indicator characteristic value in digital form, the standard field with a matching relationship, and the standardized indicator name.

[0074] The step S312 can be obtained by the following steps 1 to 3:

[0075] Step 1: Based on the standard fields with matching relationships and the standardized indicator names, standard feature values ​​are obtained from the medical knowledge database.

[0076] Step 2: Compare the indicator characteristic value in digital form with the standard characteristic value to determine the abnormal indicator characteristic value.

[0077] Step three: determine the matching result based on the abnormal indicator feature value.

[0078] The above-mentioned standard characteristic value can be a specific value or a numerical range; the above-mentioned abnormal indicator characteristic value can be understood as an indicator characteristic value in digital form that does not conform to the standard characteristic value. For example, the "anteroposterior diameter" of the "lateral lobe of the thyroid gland" in the medical record is "3.5", and the standard range indicated by the standard characteristic value in the knowledge database is "1-2". Through comparison, it is confirmed that the "anteroposterior diameter" of the "lateral lobe of the thyroid gland" in the medical record exceeds the upper limit of the standard range indicated by the standard characteristic value, which is an abnormal indicator characteristic value. Based on the abnormal indicator characteristic value, it can be concluded that the anteroposterior diameter property characteristics of the lateral lobe of the thyroid gland belong to the abnormal range.

[0079] The step three can be implemented by the following steps A to E:

[0080] Step A: based on the abnormal indicator characteristic value, determine the first inspection result corresponding to the abnormal indicator characteristic value.

[0081] Continuing with the example that the "anteroposterior diameter" of the "lateral lobe of the thyroid gland" in the medical record exceeds the upper limit of the standard range indicated by the standard characteristic value, based on the knowledge graph, the symptom that exceeds the upper limit of the normal items of the "lateral lobe of the thyroid gland" and the "anteroposterior diameter" can be matched as "thyroid thickening", which is the first examination result mentioned above.

[0082] Step B: Filter out the designated characteristic values ​​containing designated characters from the other indicator characteristic values ​​except the indicator characteristic values ​​in digital form in the structured data of the medical examination report after structured processing.

[0083] The above-mentioned designated characters can be set according to actual needs. For example, the designated characters can include "yes" and the like. In actual implementation, other indicator characteristic values ​​except for the indicator characteristic values ​​in digital form usually cannot be directly compared with the standard data in the medical knowledge database. At this time, other characteristic values ​​in the structured medical examination report can be matched with the standard data in the medical knowledge database through certain rules. Specifically, the designated characteristic values ​​containing the designated characters can be filtered out from other characteristic values, for example, the data whose value of quantization_boolean is "yes" and the like.

[0084] Step C, obtaining the designated index name and the inspection part name corresponding to the designated characteristic value.

[0085] Step D, based on the specified characteristic value, spell the specified indicator name and the inspection part name according to the preset spelling rules to obtain a spelling result; based on the spelling result, determine the second inspection result corresponding to the specified characteristic value.

[0086] The above-mentioned designated indicator names can also be understood as the names of items in the medical examination report after structured processing; in actual implementation, since the establishment of the medical case model is based on the medical knowledge specification, which is consistent with the establishment standard of data in the medical knowledge database, the nature characteristics of the examination report can be matched with the knowledge data through certain rules. The following are the specific matching rules:

[0087] Take the data whose value of quantization_boolean is "yes", and spell the value of quantization_project and the value of quantization_text corresponding to the "part name" according to the rule of "part name" + "project name" (project name without "yes"), match the spelled content with the symptoms of the knowledge base, and get the abnormal situation.

[0088] For example, the structured inspection report data is as follows:

[0089]

[0090] The value of quantization_boolean is "yes", and the quantization_text corresponding to the name of the part where quantization_project "Are the lymph nodes enlarged?" is "neck". According to the spelling rule of "neck" + "lymph node enlargement", the symptom "cervical lymph node enlargement" is matched, and the abnormal condition in the examination report is "cervical lymph node enlargement", which is the second examination result mentioned above.

[0091] Step E: determining a matching result based on the first inspection result and the second inspection result.

[0092] The step E may specifically include the following steps a and b:

[0093] Step a, obtaining designated information in an unstructured medical examination report; wherein the designated information includes at least one of the following: gender information, age information.

[0094] For an examination item, different populations, ages, body positions, pregnancy conditions, etc. may have different medical imaging manifestations, and the judgment criteria for whether the corresponding knowledge data is abnormal will also be different. Therefore, when judging whether the examination report is abnormal, it is also necessary to combine the population, age, pregnancy and other information in the medical record information, and comprehensively compare the knowledge data in the medical knowledge database to determine whether there are abnormal indicators in the examination report.

[0095] Step b: determining a matching result based on the specified information, the first check result and the second check result.

[0096] The step b may specifically include:

[0097] Step (1), screening out abnormal inspection results that meet the specified information from the first inspection result and the second inspection result.

[0098] Step (2), obtaining interpretation information matching the abnormal examination result; wherein the interpretation information includes at least one of the following: a medical explanation or health guidance suggestion corresponding to the abnormal examination result.

[0099] Step (3), determining the matching result based on the abnormality inspection result and the interpretation information.

[0100] Check the knowledge data sample as shown in Table 4:

[0101] Table 4

[0102]

[0103] First, obtain the basic information of the patient in the current medical record, such as "gender: male, age: 73 years old"; then, divide the age into groups according to the age classification standard, as shown in Table 5, and finally combine the patient information, match the knowledge data, and determine whether it is an abnormal indicator.

[0104] Table 5

[0105] crowd age Newborns Within 28 days baby 28 days to 1 year old child 0-6 years old juvenile 7-17 years old youth 18-40 years old middle aged 41-65 years old elderly After 66 years old

[0106] For example, the medical record data inspection report is: "Transrectal scan: transverse diameter of the prostate 4.3cm, upper and lower oblique diameter 4.0cm. The inner gland is 3.2x2.4cm. The inner gland is moderately to severely enlarged, the prostate contour is clear, the echo is uneven, and no obvious tumor is found."

[0107] The knowledge data is: "1. Size: The diameter of the prostate increases with age. In young people, the average superior and inferior diameter, transverse diameter and anterior-posterior diameter of the prostate are 3.0cm, 3.1cm and 2.3cm respectively, while in the elderly they are 5.0cm, 4.8cm and 4.3cm respectively. 2. Prostate enlargement is a common abnormal sign, which is manifested by the transverse diameter of the prostate being greater than 5cm or the prostate can still be seen at 2cm above the pubic symphysis."

[0108] Referring to the aforementioned quantifiable field matching rules, comparing the size of the prostate transverse diameter, the prostate transverse diameter is 4.3cm. If this report is for a young or middle-aged patient, it is an abnormal indicator, and if it is for an elderly patient, it is a normal indicator.

[0109] Step S314: Generate an interpretation file corresponding to the unstructured medical examination report based on the matching result.

[0110] The knowledge entities judged as abnormal indicators are retained, their medical explanations and health guidance suggestions in the knowledge base are read, and the examination report interpretation file is generated.

[0111] The above-mentioned medical examination report interpretation method can automatically interpret unstructured medical imaging examination reports and medical pathology examination reports, use natural language processing technology to structure the text data in the examination report, and match it with the standard data in the medical knowledge database. In addition, this solution can also combine the patient's medical record information to obtain abnormal indicators for interpreting the examination report, so that the final examination report has more personalized interpretation characteristics.

[0112] To further understand the above embodiments, the following is provided: Figure 4 The flowchart of another method for interpreting medical examination reports shown in the figure first obtains the examination report case data, structures the parts and property characteristics in the examination report, and compares them with the standard data of the medical knowledge base to determine abnormal situations. For situations that can be directly compared, the property characteristic values ​​can be standardized, and a matching relationship between the medical record field and the knowledge field can be established to obtain the abnormal situation after matching. For situations that cannot be directly compared, the property characteristics and knowledge data can be matched according to rules to obtain the abnormal situation after matching. After all the abnormal situations in the medical record data of the examination report are obtained, the abnormal indicators of the examination report can be read in combination with the patient's medical record information, and the medical explanation and health guidance significance of the abnormal indicators can be retrieved, and finally an examination report interpretation file is generated.

[0113] For patients who lack medical knowledge, when they receive medical imaging reports or pathology examination reports, they hope to understand their current abnormal indicators, medical explanations and health guidance suggestions through automatic interpretation of the reports. The above method can meet this need.

[0114] Since the writing of medical imaging and pathology examination reports in electronic medical records is different and non-standard, the interpretation needs of unstructured examination reports can be solved by structuring and standardizing unstructured examination reports and matching them with standard knowledge data in the medical knowledge base. In addition, this method can combine patient medical record information, match knowledge data, and judge abnormal indicators of examination reports, making the interpretation of examination reports more personalized.

[0115] The above-mentioned medical examination report interpretation method is to interpret the examination report according to the individual situation of each patient by comparing the fields and data in the examination report with the medical knowledge base, reading the medical explanations and health guidance suggestions corresponding to the abnormal fields. It can help patients interpret the abnormalities in the examination report according to their own situation and give specific health guidance suggestions, which can help patients quickly and directly understand their health status and reduce the waiting time for medical treatment.

[0116] The embodiment of the present invention provides a medical examination report interpretation device, such as Figure 5 As shown, the device includes: an acquisition module 50, used to acquire an unstructured medical examination report; a processing module 51, used to structure the unstructured medical examination report based on a preset medical case model and using a natural language processing method; a matching module 52, used to match the structured data in the structured medical examination report with the standard data in the medical knowledge database acquired in advance to obtain a matching result; and a generation module 53, used to generate an interpretation file corresponding to the unstructured medical examination report based on the matching result.

[0117] The above-mentioned medical examination report interpretation device obtains an unstructured medical examination report; based on a preset medical case model, the unstructured medical examination report is structured by using a natural language processing method; the structured data in the structured medical examination report is matched with the standard data in the pre-acquired medical knowledge database to obtain a matching result; based on the matching result, an interpretation file corresponding to the unstructured medical examination report is generated. The device uses a medical case model and natural language processing technology to structure the unstructured medical examination report and match it with the standard data in the medical knowledge database, and generates an interpretation file based on the matching results, which can meet the patient's demand for automatic interpretation of unstructured text-based examination reports.

[0118] Furthermore, the processing module is also used to: use natural language processing to store the unstructured data in the unstructured medical examination report according to a structure predefined by the medical case model to obtain a structured medical examination report.

[0119] Furthermore, the matching module is also used to: filter out indicator characteristic values ​​in digital form from the structured data of the structured medical examination report; match the structured fields associated with the indicator characteristic values ​​in digital form in the structured medical examination report with multiple fields in the pre-acquired medical knowledge database to obtain standard fields with a matching relationship; standardize the indicator names associated with the indicator characteristic values ​​in digital form to obtain standardized indicator names; determine the matching results based on the indicator characteristic values ​​in digital form, the standard fields with a matching relationship and the standardized indicator names.

[0120] Furthermore, the matching module is also used to: obtain standard feature values ​​from the medical knowledge database based on standard fields and standardized indicator names with matching relationships; compare the indicator feature values ​​in digital form with the standard feature values ​​to determine abnormal indicator feature values; and determine the matching results based on the abnormal indicator feature values.

[0121] Furthermore, the matching module is also used to: determine a first inspection result corresponding to the abnormal indicator characteristic value based on the abnormal indicator characteristic value; filter out a specified characteristic value containing specified characters from other indicator characteristic values ​​except the indicator characteristic values ​​in digital form in the structured data of the medical examination report after structured processing; obtain a specified indicator name and an inspection part name corresponding to the specified characteristic value; based on the specified characteristic value, spell the specified indicator name and the inspection part name according to preset spelling rules to obtain a spelling result; based on the spelling result, determine a second inspection result corresponding to the specified characteristic value; and determine a matching result based on the first inspection result and the second inspection result.

[0122] Furthermore, the matching module is also used to: obtain specified information in an unstructured medical examination report; wherein the specified information includes at least one of the following: gender information, age information; and determine a matching result based on the specified information, the first examination result and the second examination result.

[0123] Furthermore, the matching module is also used to: screen out abnormal inspection results that meet the specified information from the first inspection results and the second inspection results; obtain interpretation information that matches the abnormal inspection results; wherein the interpretation information includes at least one of the following: medical explanations and health guidance suggestions corresponding to the abnormal inspection results; and determine the matching results based on the abnormal inspection results and the interpretation information.

[0124] The medical examination report interpretation device provided in the embodiment of the present invention has the same implementation principle and technical effects as those in the aforementioned medical examination report interpretation method embodiment. For the sake of brief description, for matters not mentioned in the embodiment of the medical examination report interpretation device, reference may be made to the corresponding contents in the aforementioned medical examination report interpretation method embodiment.

[0125] The embodiment of the present invention further provides an electronic device, see Figure 6 As shown, the electronic device includes a processor 130 and a memory 131. The memory 131 stores machine executable instructions that can be executed by the processor 130. The processor 130 executes the machine executable instructions to implement the above-mentioned medical examination report interpretation method.

[0126] Further, Figure 6 The electronic device shown further includes a bus 132 and a communication interface 133 , and the processor 130 , the communication interface 133 and the memory 131 are connected via the bus 132 .

[0127] The memory 131 may include a high-speed random access memory (RAM), and may also include a non-volatile memory, such as at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 133 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used. The bus 132 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 6 Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0128] The processor 130 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 130. The above processor 130 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present invention can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present invention can be directly embodied as a hardware decoding processor for execution, or a combination of hardware and software modules in the decoding processor for execution. The software module may be located in a storage medium mature in the art, such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory, or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 131, and the processor 130 reads the information in the memory 131 and completes the steps of the method of the above embodiment in combination with its hardware.

[0129] An embodiment of the present invention also provides a machine-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the above-mentioned medical examination report interpretation method. The specific implementation can be found in the method embodiment, which will not be repeated here.

[0130] The computer program products of the medical examination report interpretation method, device and electronic device provided in the embodiments of the present invention include a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the methods described in the previous method embodiments. The specific implementation can be found in the method embodiments, which will not be repeated here.

[0131] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for interpreting a medical examination report, characterized in that: The method comprises: Obtain unstructured medical examination reports; Based on a preset medical case model, the unstructured medical examination report is structured by using a natural language processing method; wherein the medical case model is a structured data structure defined based on electronic medical record data and according to applicable scenarios; Matching the structured data in the structured medical examination report with the standard data in the pre-acquired medical knowledge database to obtain a matching result; Based on the matching result, generating an interpretation file corresponding to the unstructured medical examination report; The step of matching the structured data in the structured medical examination report with the standard data in the pre-acquired medical knowledge database to obtain a matching result includes: Filtering out an indicator characteristic value in digital form from the structured data of the medical examination report after structured processing; wherein the indicator characteristic value is a property characteristic value in the medical examination report; Matching the structured fields associated with the numerical indicator characteristic values ​​in the structured medical examination report with the plurality of fields in the pre-acquired medical knowledge database to obtain standard fields with matching relationships; Standardizing the indicator name associated with the indicator characteristic value in digital form to obtain a standardized indicator name; wherein the indicator name is the name of the property characteristic in the medical examination report; The matching result is determined based on the indicator characteristic value in digital form, the standard field having a matching relationship, and the standardized indicator name.

2. The method according to claim 1, characterized in that The step of structurally processing the medical examination report by using a natural language processing method based on a preset medical case model includes: By adopting a natural language processing method, the unstructured data in the unstructured medical examination report is stored according to a structure predefined by the medical case model, so as to obtain the structured medical examination report.

3. The method according to claim 1, characterized in that The step of determining the matching result based on the indicator characteristic value in digital form, the standard field having a matching relationship and the standardized indicator name comprises: Based on the standard fields and the standardized indicator names having a matching relationship, obtaining standard feature values ​​from the medical knowledge database; Compare the indicator characteristic value in digital form with the standard characteristic value to determine the abnormal indicator characteristic value; Based on the abnormal indicator characteristic value, the matching result is determined.

4. The method according to claim 3, characterized in that The step of determining the matching result based on the abnormal indicator characteristic value comprises: Based on the abnormal indicator characteristic value, determining a first inspection result corresponding to the abnormal indicator characteristic value; From the structured data of the medical examination report after structured processing, filter out the specified characteristic values ​​containing the specified characters from other indicator characteristic values ​​except the indicator characteristic values ​​in the digital form; Obtaining the designated indicator name and the inspection part name corresponding to the designated characteristic value; Based on the designated characteristic value, the designated indicator name and the inspection part name are spelled according to a preset spelling rule to obtain a spelling result; based on the spelling result, a second inspection result corresponding to the designated characteristic value is determined; The matching result is determined based on the first inspection result and the second inspection result.

5. The method according to claim 4, characterized in that The step of determining the matching result based on the first inspection result and the second inspection result comprises: Acquire designated information in the unstructured medical examination report; wherein the designated information includes at least one of the following: gender information, age information; A matching result is determined based on the specified information, the first check result, and the second check result.

6. The method according to claim 5, characterized in that The step of determining a matching result based on the specified information, the first check result and the second check result comprises: Filtering out abnormal inspection results that meet the specified information from the first inspection result and the second inspection result; Obtaining interpretation information matching the abnormal test result; wherein the interpretation information includes at least one of the following: medical explanations and health guidance suggestions corresponding to the abnormal test result; The matching result is determined based on the abnormality inspection result and the interpretation information.

7. A medical examination report interpretation device, characterized in that: The device comprises: An acquisition module, used to acquire unstructured medical examination reports; A processing module, used for structuring the unstructured medical examination report by using a natural language processing method based on a preset medical case model; wherein the medical case model is a structured data structure defined based on electronic medical record data and according to an applicable scenario; A matching module is used to match the structured data in the medical examination report after structured processing with the standard data in the medical knowledge database acquired in advance to obtain a matching result; A generating module, configured to generate an interpretation file corresponding to the unstructured medical examination report based on the matching result; The matching module is also used for: Filtering out an indicator characteristic value in digital form from the structured data of the medical examination report after structured processing; wherein the indicator characteristic value is a property characteristic value in the medical examination report; Matching the structured fields associated with the numerical indicator characteristic values ​​in the structured medical examination report with the plurality of fields in the pre-acquired medical knowledge database to obtain standard fields with matching relationships; Standardizing the indicator name associated with the indicator characteristic value in digital form to obtain a standardized indicator name; wherein the indicator name is the name of the property characteristic in the medical examination report; The matching result is determined based on the indicator characteristic value in digital form, the standard field having a matching relationship, and the standardized indicator name.

8. An electronic device, characterized in that: It includes a processor and a memory, wherein the memory stores machine executable instructions that can be executed by the processor, and the processor executes the machine executable instructions to implement the medical examination report interpretation method according to any one of claims 1 to 6.

9. A machine-readable storage medium, characterized in that: The machine-readable storage medium stores machine-executable instructions. When the machine-executable instructions are called and executed by the processor, the machine-executable instructions prompt the processor to implement the medical examination report interpretation method described in any one of claims 1-6.

Images