• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    SYSTEM AND METHOD OF AID TO CREATE REPORTS, WORKSTATION AND COMPUTER PROGRAM PRODUCT A system of aid to report creation is revealed. The system comprises a set of associations (1), in which an association associates at least a first concept term in a knowledge domain and at least a second concept term in a knowledge domain, in which the association indicates a frequent co-existence. occurrence of at least one first term of concept and at least one second term of concept in reports in the knowledge domain. The system comprises a concept term extractor (2) to extract at least a first concept term relating to a specific individual from a data record (7) associated with the specific individual, the data record (7) comprising a report at least partially completed (5) related to the specific individual. The system comprises a missing concept term searcher (3) to find at least one first concept term (6) based on the set of associations (1), where at least one second concept term (11) is missing at least partially completed report (5).
    公开号:BR112013002534B1
    申请号:R112013002534-4
    申请日:2011-08-05
    公开日:2020-12-22
    发明作者:Michael Chun-Chieh Lee；Eric Cohen-Solal
    申请人:Koninklijke Philips N.V；
    IPC主号:
    专利说明:

    FIELD OF THE INVENTION
    The invention concerns the aid in creating reports. HISTORY OF THE INVENTION
    A Radiology Information System (RIS) integrated with Image Archiving and Communication System (PACS) capabilities helps hospitals and image acquisition centers to optimize their flow and manage the circulation of information and images throughout the facility to provide efficient healthcare and services to the patient. The different steps involved in the present process start with a doctor requesting a radiology procedure, a patient going to the radiology department, the image acquisition procedure being carried out, a radiologist reading the images and preparing a report to be sent to the requesting doctor .
    The report that is produced by the radiologist usually summarizes the relevant aspects of the patient's history and the clinical question being asked and then indicates which regions of the radiology image were examined and what were the findings, if any. Findings and any conclusions or impressions that can be drawn from the results are usually given in separate sections of the report. In general, the requesting physician expects the findings and impressions to address the clinical issue that he or she initially asked. If the findings and impressions inadequately address the clinical issue, it may be necessary for the requesting physician to contact the radiologist and ask for clarification.
    The document Geis, J. (2 O O7), "Medical Imaging Informatics: How It Improves Radiology Practice Today," Journal of Digital Imaging 2 O (2): 99, reveals that radiologists are under pressure to add more values to the medical image acquisition - to provide more educated, accurate, useful and efficient interpretations in the face of increasingly broad and complex image acquisition studies and to communicate this information quickly and in the most useful way. The radiology department and the radiologist both need to be better, faster and cheaper. In addition, the radiologist synthesizes what he sees in the images, the clinical data, and his medical knowledge to produce the interpretation. This should be a coherent, accurate and useful debate that adds value to the images.
    Therefore, there is a need for tools that can help them deliver the desired quality in reports and still address the local restrictions of each radiology institution. There is a need for consistency among radiologists in the way they report similar cases (inter-radiologist consistency) as well as a need for consistency over time for each individual radiologist (intra-radiologist consistency). In addition, the nature of their work imposes time constraints and reading quotas, increasing the number of potentially incomplete and therefore potentially inaccurate reports. This inaccuracy will possibly lead to unnecessary interactions with requesting physicians requesting further clarification, causing interruption and decreased efficiency, which can be avoided. SUMMARY OF THE INVENTION
    It would be advantageous to obtain improved help for reporting. To better address this problem, a first aspect of the invention provides a system comprising a set of associations, in which an association associates at least one first concept term in a knowledge domain with at least one second concept term in the knowledge domain, where the association indicates a frequent co-occurrence of at least a first concept term and at least a second concept term in reports in the knowledge domain; a concept term extractor for extracting at least a first concept term related to a particular individual from a report at least partially completed related to the specific individual; a missing concept term searcher to find at least one second concept term associated with at least one first concept term based on the set of associations, where at least a second concept term is missing from the report at least partially completed ; an indicator to indicate at least a second concept term found to a user. The system thus provides an indication of missing second concept terms that are associated with first concept terms related to the individual. When providing this indication of missing second concept terms, the user is reminded to include a comment involving the second concept term in the report. In this way, it is easier to create a complete report, because any potentially missing concept terms are indicated to the user. Consequently, the number of incomplete reports can be reduced. In addition, when the report is read by a report consumer, such as a requesting physician, less interaction with the report author is necessary because the report contains information about the relevant concept terms. Alternatively, when the concept has already been discussed in the report using different terminology, the user may be reminded to reformulate the report to use the second concept term. This improves the consistency of the terminology used in the reports. The system user can understand an author of the report at least partially completed. The knowledge domain can comprise a medical knowledge domain, the individual can understand a patient, and the report can comprise a medical diagnosis or treatment report. In an effort to help ensure that medical reports adequately address the clinical question posed, for example, by a requesting physician, the system can provide a means of verifying that a radiologist or physician will complete the report using expected terminology based on the clinical context . The concept term extractor can be arranged to extract at least a first concept term from the report at least partially completed. The first concept term (s) may be present in the report at least partially completed, for example, in an introductory section in which the context of the report is defined. Such sections may include a description of the medical history and / or a description of the questions asked by a requesting person. The concept term extractor can be arranged to extract at least a first concept term from a section of the report at least partially completed. In addition, the set of associations can comprise associations by linking the first concept terms of the first section to the second concept terms belonging to a second section different from the reports in the domain of knowledge. In this way, the concepts that need to be discussed in the second section of the report, based on the first section of the report, can be identified. The concept term extractor can be arranged to extract at least one first concept term from the medical history information represented by the data record. Medical history information can provide clues as to which concept terms should be discussed in the rest of the report. The at least one second concept term can be associated with a specific section of the report, and the indicator can be arranged to indicate which section the at least one second concept term is associated with. This makes the suggestion of missing concept terms more specific, as the missing concept terms are presented in the context of the section in which they should appear. The system can comprise a report entry to receive a completed report, where the report at least partially completed is the completed report. The completed report can therefore be verified for missing concept terms. In this way, the completed report can be verified. If any missing concept terms are identified, the completed report and the missing concept terms can be presented to the author (or another author), who can then consider providing additional text involving the missing concept. The system can comprise a report creation tool to allow a user to create at least part of the report. The indicator can be arranged to indicate at least a second concept term found during a process of creating at least part of the report. For example, the indicator can be part of the reporting tool. After entering a word, the indicator can update the indication of the missing concept terms. This helps the user to verify the completeness of the report when creating the report, which helps to create reports containing the necessary information and can avoid any iterations of correcting or extending the report. At least a second medical term may comprise an anatomical region term. The system may comprise an image viewer configured to display a visualization of an individual's anatomical region based on an anatomical region term, where the anatomical region corresponds to the anatomical region term. In this way, the number of manual interactions is reduced, as the relevant anatomical region is visualized automatically. The system may include a second concept term classifier to classify at least one second concept term by type of concept, and where the indicator is arranged to indicate the concept type for a second concept term. This provides a more organized view of the concepts.
    Associations can have different intensities. The system can comprise a second concept term orderer to classify the second concept terms found by the missing concept term searcher, based on the intensities of the associations by which the second concept terms were found. In this way, the terms of concept that are most likely to be relevant for discussion in the report can be indicated as such for the user. The system can comprise an association generator to create an association based on the co-occurrence frequencies of the first and second concept terms in a set of reports. In this way, the set of associations can be generated automatically. In addition, the association set can be updated, taking into account new reports when added to the set. The system can be incorporated into a workstation, for example, a medical workstation.
    In another aspect, the invention provides a method of assisting in the creation of reports, comprising extracting at least a first term of concept related to a specific individual from a report at least partially completed related to the specific individual; the discovery of at least a second concept term associated with at least one first concept term based on a set of associations, where at least a second concept term is absent from the report at least partially completed, where an association the set of associations associates at least one first concept term in the knowledge domain with at least one second concept term in the knowledge domain, where the association indicates a frequent co-occurrence of at least one first concept term and at least a second concept term in knowledge reports; and the indication of at least a second concept term found to a user.
    In another aspect, the invention provides a computer program product comprising instructions for making a processor system perform the established method.
    It will be appreciated by those skilled in the art that two or more of the aforementioned accomplishments, implementations and / or aspects of the invention can be combined in any form considered useful. Modifications and variations of the workstation, method, and / or computer program product, which correspond to the described modifications and variations of the system, can be performed by a person skilled in the art based on this description. BRIEF DESCRIPTION OF THE DRAWINGS
    These and other aspects of the invention will become evident and will be elucidated with reference to the achievements described below. In the drawings, Figure 1 is a diagram of a reporting aid system; Figure 2 is a flowchart of a method to aid in the creation of reports; and Figure 3 is a diagram of a reporting aid system. DETAILED DESCRIPTION OF ACHIEVEMENTS
    Figure 1 illustrates a reporting aid system. The system can, in particular, be used as an aid in creating reports in a given knowledge domain. Examples of such a domain of knowledge include radiology, cardiology, oncology. It is possible to adapt the system to a more restricted domain of knowledge such as diagnostic reports of lung radiography, or to a wider domain of knowledge, such as medical reports in general. Other non-medical knowledge domains can also be applied. The figure shows an exemplary achievement with a number of characteristics. These characteristics are only described by way of example; most features are optional or can be implemented in different ways. The system can include a set of associations 1. Each association in the set can associate at least one first concept term in a knowledge domain with at least one second concept term in the knowledge domain. Associations may indicate that at least one first concept term often co-occurs with at least one second concept term in knowledge reports. In other words, reports containing the first term of concept often also contain the second term of the concept. The set of associations 1 can be prepared in advance. It is also possible to update the set of associations 1 by analyzing a set of reports, as described below. The concept terms that are used in the knowledge domain can be organized into a set of concept terms 16. Associations can also indicate that when a data record 7 contains a first concept term, the report often contains a second term concept. The system can additionally comprise a concept term extractor 2 for extracting at least a first concept term 6 from a data record 7. Data record 7 can be part of a database with data records, where each data record can contain information related to a specific individual, such as a patient. Data record 7 can be understood to also include any individual-related data files currently being edited, such as a partially completed diagnostic report that is loaded into a report editor. Therefore, the data record 7 may comprise a report at least partially completed 5 related to the specific individual. The concept term extractor may be arranged to extract concept terms corresponding to one of the concept terms in the concept term set 16. In particular, the concept term extractor 2 may be arranged to extract any concept terms that are used as first terms of concept in the association set 1. The system can additionally comprise a searcher of missing concept terms 3, to find any concept terms that are missing from the report at least partially completed, but which would be expected to appear in the report based on associations and concept terms extracted from data record 7. For this purpose, the searcher for missing concept terms 3 can identify one or more concept terms that are linked to at least one first concept term 6 extracted by the concept term extractor 2 through at least one of the associations 1. In addition, the term search engine missing concept 3 can check if any of these one or more concept terms are missing from the report at least partially completed 5. The missing concept terms are sent by the search engine for missing concept terms such as at least a second concept term 11. The The system may additionally comprise an indicator 4 to indicate the at least one second concept term found 11 to a user. This indication can be done in several ways. For example, the at least one second concept term 11 can be shown on a list screen. This can allow the user to update the report at least partially completed 5 with information including at least a second concept term 11.
    As discussed above, in medical applications, the knowledge domain may comprise a medical knowledge domain, the individual may comprise a patient, and the report may comprise a medical diagnosis or treatment report. The concept term extractor 2 can be arranged to extract at least a first concept term 6 from the at least partially completed report 5. Consequently, it is possible that the concept term extractor 2 only evaluates the report at least partially completed, not the rest of the data record 7. This may be the case when the reports are independent, for example, they contain the clinical history, all the questions asked by the requesting physician, the findings, and the diagnosis. These items can be organized into different sections of the report. Another division into sections can also be provided.
    Such systematic sectioning of reports can be used by the system. The concept term extractor 2 can be arranged to extract at least a first concept term 6 from a first section of the report at least partially completed 5. The set of associations 1 can comprise associations linking first concept terms belonging to the first section to the second terms of concept belonging to a second section different from reports in the field of knowledge. Consequently, the searcher for missing concept terms 3 can then find missing concepts in the second section of the report based on the concept terms extracted from the first section of the report. For example, the first section can comprise clinical history information and / or a diagnostic question posed, and the second section can comprise clinical findings and / or the diagnosis. The at least one second concept term 11 can be associated with a specific section of the report, in the sense that the at least one second concept term often appears in that specific section of reports in the knowledge domain. In such a case, indicator 4 may be arranged to indicate to which section the at least one second concept term is associated. For example, at least a second concept term can be displayed while the user is editing the relevant section of the report. When the user is editing another section, the second concept terms associated with that other section can be displayed. For example, when editing the "Findings" section, the second concept terms associated with the "Findings" section can be displayed; while editing the "Impressions" section, the second concept terms associated with the "Impressions" section can be viewed. The concept term extractor 2 can be arranged to extract at least one first concept term 6 from the medical history information 8 represented by the data record. This medical history information 8 can include past reports, drug use, laboratory, and others. The system may additionally comprise a user interface element 9 to allow a user to indicate that report 5 is complete, and where indicator 4 is arranged to indicate at least a second concept term found 11 in response to a user indication that report 5 is complete. The concept term extractor and the missing concept term finder can also be activated by the user's indication, but they can also work in the background while the report is being written. The indication of the missing concept terms for the user can only be given when the report 5 is considered completed, to avoid any kind of biased reports by the user. The system may additionally comprise a report creation tool 10 O to allow a user to create report 5, or at least part of it. For example, an editor can be provided, which allows a user to type (parts of) the report. The report creation tool 1 O can also be based on speech recognition. Tool 10 O can allow a user to dictate the report and extract at least one first concept term 6 from the dictated text. In addition, the missing concept terms searcher 3 can be arranged to verify that the relevant concept terms are missing from the dictated text. Indicator 4 can be arranged to indicate at least a second concept term found 11 during a process of creating at least part of the report 5. An immediate response can guide the user to the relevant concepts. The at least a second concept term 11 may comprise an anatomical region term, such as lungs or heart. The system may additionally comprise an image viewer 15 configured to indicate a view of the anatomical region corresponding to the anatomical region term. In particular, a medical image of the patient can be shown with an indication of the anatomical region. For example, the anatomical region can be indicated by highlighting or using an arrow. The system may comprise a classifier of second concept terms 12 to classify the at least one second concept term 11 by the type of concept. Examples of types of concepts can include anatomical regions, diseases, findings, diagnosis. Indicator 4 can be arranged to indicate the concept type for a second concept term 11. For example, the concept type can be displayed next to the concept term. Alternatively, the second concept terms can be ordered in groups, each group corresponding to a type of concept.
    The associations in the set of associations 1 can have different intensities. The system may additionally comprise a classifier of second concept terms 13 to classify at least one second concept term 11 by the searcher for missing concept terms 3. The classification allows the user's attention to be drawn to the most relevant missing concept terms . The classification can be based on the intensities of the associations by which the second 11 concept terms were found. The system can additionally comprise an association generator 14 to create an association based on the co-occurrence frequencies of the first and second concept terms in a set of reports. These co-occurrence frequencies can be used to estimate the probability that a report containing a first concept term will also contain a second concept term. When this probability is high enough, the association linking these concept terms can be stored in association set 1. This will be explained below. Other ways to obtain the association set 1 are also possible. For example, associations could be created manually. Associations can also be derived from the co-occurrence frequencies of the first and second concept terms in a set of textbooks or academic articles on the domain, for example, through data mining by co-occurrence frequencies in journals doctors. This could be done together or in place of the analysis of the set of reports. The established system can be implemented as a workstation loaded with a suitable application program.
    Figure 2 illustrates a method of assisting the creation of reports. The method may comprise step 2 O1 of extracting at least a first concept term related to a specific individual from a data record associated with the specific individual, the data record comprising a report at least partially completed related to the specific individual . The method may additionally comprise step 2 O2 of discovering at least one second concept term associated with at least one first concept term based on a set of associations, where at least one second concept term is absent from the report by least partially concluded, in which an association of the set of associations associates at least a first concept term in the knowledge domain with at least a second concept term in the knowledge domain, where the association indicates a frequent co-occurrence of at least one first concept term and at least a second concept term in knowledge reports. The method additionally comprises step 2 O3 of indicating at least one second concept term found to a user. The method can be changed or extended, for example, based on the functionality described in this document in relation to the system. The method can be implemented as a computer program product comprising instructions for making a processor system perform the method.
    In addition to radiology reports, the system can be used in other medical areas for which reports are routinely generated. Examples include, but are not limited to, cardiology, oncology, pathology, or surgery. As a concrete example for cardiology, a report for which "pericarditis" is the diagnosis can be reasonably expected to mention "attrition" as one of the findings. A report without this term can be flagged, since it is without a term that the reader can expect (that is, the finding must be mentioned as being present or absent) or the terminology actually used in the report is, somehow, outside standard, and a more consistent use of terminology is appropriate.
    One of the applications of the techniques described in this document is the provision of functionality to discover and suggest relevant or missing medical concept terms to possibly be added to the report by examining previous reports, being a source of an enormous amount of information on medical practices. The approach to creating a set of associations 1 may include extracting key associations or correlations between the different pieces of information contained in previous reports. The key correlations, which are used in this approach, can understand the correlations between the patient's clinical history (location (s) of the body, symptoms, signs, reason (s) for the examination, prior knowledge) and the findings of radiologists, the reported anatomical regions, impressions and diagnoses. Other significant correlations between different findings within the report or between findings and diagnoses can also be identified and used.
    An example of such a correlation is as follows. Consider a CT scan of a patient with a "tinnitus" complaint. Analysis of the reports may show that it is customary for a radiologist to describe the patient's "ear canals". In some cases, this may be reporting that there are no problems in this area, while in others, a specific pathology is reported. However, a doctor who has ordered this CT can reasonably expect to receive the CT report and find some indication of the status of the patient's "ear canals".
    The indication of the missing concept terms can be used in a situation where the radiologist is about to present the patient's case report and asks the system to perform a verification analysis (post-report verification). The suggestion of possibly relevant medical terms is also anticipated earlier, during the report generation process itself, while the radiologist is typing or dictating a report (with speech-to-text technology), or even earlier while reading, the interpretation process, where recommendations can be suggested. As an example, offering recommendations to a resident in training can provide access to knowledge that the resident does not yet know or have.
    The following are brief descriptions of characteristics in the context of an example of a predicted situation:
    Post-report verification: Suggest medical terms or concepts to improve the clinical quality of the report when the report is complete. In this situation, the objective is to provide relevant and absent terms to address the consistency, completeness and accuracy of each report produced. Given the clinical history information (body location (s), symptoms, signs, reason (s) for the examination, prior knowledge) for the current patient study, the system can suggest medical concept terms that are typically associated with the history this specific patient, but are not found in the final report. The radiologist can then read the suggestions and add terms to the report, if necessary. The missing terms can be medical concepts related to the findings, anatomical regions, impressions and diagnoses. In particular, consistency is not necessarily driven by establishing a set of rules on reporting practices (for a particular institution, for example) that physicians must comply with. This is done through a tool that perfectly suggests medical terms without having to remember which rules apply to this specific patient case.
    Reporting assistance: Suggest medical terms when creating the patient report. While the doctor is generating the report, medical terms can be displayed, which are known to be associated with the available medical history terms. At the beginning of the report, a list of terms is suggested. The radiologist can decide to look at these terms or not. While he is writing or dictating the report, more input terms become available to the system, allowing suggestions of medical terms more centered on the context (correlations between findings or between findings and diagnoses are used). It is also possible to provide an automatic completion function, since the simple writing of the beginning of a word drastically reduces the possible terms to be proposed. In such a case, the system may have more influence on the selection of terms, but it offers the possibility to better clarify the description and, perhaps, give a tip to the radiologist to start an investigation course not considered yet.
    Region adviser through intelligent classification during reading: The radiologist often knows which regions of the image to investigate to look for specific and abnormal visual patterns or characteristics, given the clinical history. One of the purposes may be to promote the terms of the most likely region, based on what has been reported in previous patient studies for a similar clinical history, and consequently to prevent regions from being overlooked. It also promotes the unnoticed discovery of new associations between, for example, a symptom and an anatomical region, which were not necessarily taught in books or in medical training. Association set 1 can be created based on an analysis of previous reports. Such analysis may include the stage of extracting relevant pieces of information from radiology reports. Radiology reports are artifacts and the end result of the radiology reading exercise. A radiology report contains the patient's clinical history (location (s) of the body, symptoms, signs, reason (s) for the examination, prior knowledge) and the radiologist's findings, the relevant anatomy sites and the conclusion with the diagnosis current patient. The report can range from well structured (form) to less structured (most cases today) with sections containing free text.
    In the case of free text reports, existing natural language processing (NLP) techniques can be used to automatically extract the relevant pieces of information in the form of medical concepts (or terms).
    As an illustrative example of the clinical history: "Patient with a history of neurofibromatosis and hypothalamic astrocytoma. Patient with severe headaches on the right side where the anastomosis is located." In this case, the system can identify four main medical history terms: astrocytoma, headaches, anastomosis and neurofibromatosis. From the same report, the system can extract medical concepts (terms) from both the finding sections and the concluding sections: basal ganglia, dentate core, frontal lobe, hypothalamus, nerve, optic chiasm, pellucid septum, soft tissues, third ventricle .
    The analysis of previous reports may additionally comprise the identification of associations and correlations of medical terms. At this stage, relevant statistics can be collected. These statistics can comprise the frequency of finding a given medical concept within all reports and / or the frequency of co-occurrence between two different types of medical concepts (a co-occurrence being defined as two different terms appearing both at least once in a single report). For example, the frequency with which a concept term appears in the clinical history can be determined as a term frequency, and the frequency with which a first concept term in a clinical history co-occurs with a second concept term in a finding / diagnosis can be determined as a co-occurring frequency. Therefore, the count of co-occurrences can be restricted based on the location of the report's terms. For example, a co-occurrence between term A and term B can be counted only if term A appears in the "History" section of the report and term B appears only in the "Findings" section of the report. However, this is not a limitation. The count of co-occurrences can be weighted by the fact that a positive diagnosis has been detected. In addition, the co-occurrence count can be weighted by the seniority of the reading radiologist.
    Analysis of previous reports may additionally comprise the selection of associations and correlations of significant medical terms. Some of the correlations identified may be insignificant or simply random. In an exemplary embodiment, a Fisher's exact test was applied and the p-value obtained is used to classify the list of associations based on their significance, ranging from the most significant to the most likely random at the bottom of the list. Other statistical tests, such as Chi-square tests, are also applicable.
    As an example, consider a set of 8 O reports, in which the term A appears at least once in every 7 O reports, and the term B appears at least once in every 10 O reports. If these terms appear randomly in the reports (no real significant association), then, on average, about 2 of the 8 OO reports would be expected to contain both the term A and the term B. If, instead, 8 is observed, or 9, or 1 O reports with both terms, so there may be a good chance that this is a real significant association. A variety of well-known statistical tests are available to discover this situation in a rigorous manner. The choice of a specific test is not intended to limit the invention.
    Analysis of previous reports may additionally comprise the estimation of conditional probabilities linking medical terms. In an exemplary embodiment, significant associations are found using conditional probabilities. The probabilities of interest here may include: P (finding term | clinical history term), and / or P (diagnostic term | clinical history term). More generally, the conditional probability can be formulated as: P (second concept term | first concept term). A conditional probability can be estimated by taking the relationship between the frequency of co-occurrence and the frequency of the first concept term using the Bayes equation, as follows: P (second concept term | first concept term) = P ( first concept term AND second concept term) / P (first concept term). The frequency of co-occurrence can be used to estimate P (first concept term AND second concept term). The frequency of occurrence of the first concept term can be used to estimate P (first concept term). The same principle can be applied to calculate P (finding term | clinical history term) and / or P (diagnostic term | clinical history term).
    It is also possible to estimate multi-term conditional probabilities, such as P (finding terms | clinical history term a, clinical history term b) or higher order. For the sake of simplicity, the rest of the description will use single term probabilities, but a higher order probability can be applied in a similar way.
    When analyzing a current report, recommendations can be generated using the set of associations 1. At the same time, the set of associations 1 can be updated by the association generator 14, for example, by updating the conditional probabilities discussed above.
    Several different usage regimes can be implemented to indicate a suggestion of missing concept terms.
    In a first usage regime, post-report verification is used. Missing concept terms are indicated after the report has been completed. If missing concept terms are found, the user can decide to update the completed report. For example, for each first concept term (or a combination of first concept terms), the system can use the pre-calculated list of significant associations to find a relevant second term or concept terms (s). A 1 O threshold can be determined in advance, or chosen by the user (for example, by interaction through a slider mechanism) to limit the list of associations. The completed report can then be analyzed to extract the concept terms contained using 15 NLP techniques. This list of terms in the report is compared with the terms corresponding to the significant associations to display only the missing terms as suggestions. The explanation of why a particular term is suggested can also be indicated to the user, for example, showing the 2 The first concept term and the intensity of the association.
    The displayed list of suggested terms can be classified based on the calculated conditional probabilities, or based on a differently calculated measure of association strength. Post-report verification 25 can be done right after the report is completed or offline. In the latter case, reports flagged with missing terms may later be brought to the attention of the radiologist or selected for peer review. Other types of correlations can be identified, such as those between two finding concepts or between a finding concept and a diagnostic concept. These concepts were produced by the radiologist and bring important contextual information that could be used to add more weight to the medical concept terms and to reduce the list of suggested terms. As an example, if a suggested term is a "a" finding and if another "b" finding is already in the report, a high probability P (finding term "a" | finding term "b") may promote the finding "a" "to the top of the list of suggested terms.
    In a second usage regime, report assistance is provided by suggesting medical terms when creating the report. This situation is different from the previous one, since the suggested terms are offered in real time when the radiologist is typing or dictating the report. This gives the doctor the possibility to return to images for better analysis. This functionality can be used, for example, for training purposes, either for resident doctors or for continuing medical education.
    In a third usage regime, a region recommender based on image viewer 15 helps the user to quickly find the relevant portions of the image corresponding to the second missing concept terms. Based on the first concept terms, for example, extracted from a clinical history that leads to the image exam, the user can be presented with a list of regions using the method described above. The interaction with the image can include the use of image processing techniques to segment the image (mesh model) and an atlas to map each region with the corresponding region concepts. As a result, by clicking on one of the suggested regions in the list (the list can be ordered by relevance using the conditional probabilities), the corresponding region can be highlighted.
    Figure 3 illustrates another example of a reporting aid system. The components of this system are largely similar to the system illustrated in Figure 1. Previous reports from a 3 O1 database can be analyzed using a 3 O2 extraction subsystem that converts the raw text in terms of extracted 3 O3 reports. This extraction 3 O2 subsystem can be similar to the concept term extractor 2, and organized to be applied to previous reports from the 3 O1 database. The 3 O4 correlation discovery subsystem checks the report terms extracted 3 O3 from a large number of reports in the database and identifies significant associations. These can be stored conceptually in a database of significant correlations 3 O5, similar to the set of associations 1. The correlation discovery subsystem 3 O4 can be similar to the association generator 14. When a new report 3 O6 (similar to the report at least partially completed 5) is being considered, the same or a different extraction subsystem 3 O7 (which may be similar to the concept term extractor 2) is executed to convert the raw report 3 O6 (for example, patient history ) in terms of extracted report 3 O8. These 3 O8 extracted report terms can be similar to at least one first concept term 6. These 3 O8 extracted report terms are inserted in a 3 O9 recommendation subsystem (which may be similar to the missing 3 concept terms searcher) which uses the associations or correlations of discovered terms 3 O5 to produce a list of new recommended terms 31 O, which can be similar to at least a second concept term 11. These new recommended terms 31 O can be displayed through a subsystem of display 311 (which may be similar to indicator 4) when applying one or more of the features described above. Note that the various subsystems and databases mentioned in this description can be part of a single physical computer system or multiple physical computer systems.
    It will be appreciated that the invention also applies to computer programs, especially computer programs on or in a carrier, adapted to put the invention into practice. The program can be in the form of a source code, an object code, an intermediate code, source code and object as in a partially compiled form, or in any other form suitable for use in implementing the method according to the invention. . It will also be appreciated that such a program can have many different architectural designs. For example, a program code implementing the functionality of the method or system according to the invention can be subdivided into one or more subroutines. Many different ways of distributing the functionalities among these subroutines will become evident to the technician in the subject. Subroutines can be stored together in an executable file to form an independent program. Such an executable file may comprise computer executable instructions, for example, processor instructions and / or interpreter instructions (for example, Java interpreter instructions). Alternatively, one or more or all of the subroutines can be stored in at least one external library file and associated with a main program statically or dynamically, for example, at run time. The main program contains at least one call to at least one of the subroutines. Subroutines can also include function calls to each other. An embodiment related to a computer program product comprises computer executable instructions corresponding to each processing step of at least one of the methods set forth herein. These instructions can be subdivided into subroutines and / or stored in one or more files that can be associated statically or dynamically. Another realization related to a computer program product comprises computer executable instructions corresponding to each medium of at least one of the systems and / or products established herein. These instructions can be subdivided into subroutines and / or stored in one or more files that can be associated statically or dynamically. The holder of a computer program can be any entity or device capable of carrying the program. For example, the carrier can include a storage medium, such as a ROM memory, for example, a CD-ROM or a semiconductor ROM memory, or a magnetic recording medium, for example, a hard disk. In addition, the carrier can be a transmissible carrier, such as an electrical or optical signal, which can be carried via an electrical or optical cable or by radio or other means. When the program is carried out on such a signal, the carrier may consist of such a cable or other device or means. Alternatively, the carrier can be an integrated circuit in which the program is embedded, the integrated circuit being adapted to carry out, or be used to carry out the relevant method. It should be noted that the above-mentioned achievements illustrate, rather than limit, the invention, and that those skilled in the art will be able to design many alternative embodiments, without departing from the scope of the appended claims. In the claims, any reference signs placed in parentheses should not be construed as limiting the invention. The use of the verb "to understand" and its conjugations does not exclude the presence of elements or steps other than those indicated in a claim. The article "one" or "one" preceding an element does not exclude the presence of a plurality of such elements. The invention can be implemented by means of hardware comprising several distinct elements, and by means of a properly programmed computer. In the device claim enumerating several means, several of these means can be realized by the same piece of hardware. The mere fact that certain measures are cited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage.
    权利要求:
    Claims (11)
    [0001]
    1. SYSTEM OF ASSISTANCE TO CREATE REPORTS, characterized by understanding; - a set of associations (1), where an association associates at least a first concept term in a knowledge domain with at least a second concept term in a knowledge domain, where the association indicates frequent co-occurrence at least a first concept term and at least a second concept term in knowledge reports; - a concept term extractor (2) to extract at least a first concept term (6) related to a specific individual from a report at least partially completed (5) related to the specific individual; - a searcher for missing concept terms (3) to find at least one second concept term (11) associated with at least one first concept term (6) based on the set of associations (1), where at least a second concept term (11) is absent from the report at least partially completed (5); - an indicator (4) to indicate at least a second concept term (11) found for a user, and a report creation tool (1 O) to allow a user to create at least part of the report (5), and where indicator (4) is arranged to indicate at least a second concept term (11) found during a process of creating at least part of the report (5), and where at least a second concept term ( 11) comprises an anatomical region term, and the system additionally comprises an image viewer (15) configured to indicate a view of an individual's anatomical region based on the anatomical region term, where the anatomical region corresponds to the term anatomical region.
    [0002]
    2. SYSTEM, according to claim 1, wherein the domain of knowledge is characterized by comprising a domain of medical knowledge, the individual comprises a patient, and the report comprises a report of medical diagnosis or treatment.
    [0003]
    3. SYSTEM, according to claim 1, in which the concept term extractor (2) is arranged to extract at least a first concept term (6) from a first section of the report at least partially completed (5), and in which the set of associations (1) is characterized by comprising associations that link first concept terms belonging to the first section to second concept terms belonging to a different second section of reports in the domain of knowledge.
    [0004]
    4. SYSTEM, according to claim 1, in which the concept term extractor (2) is arranged to extract at least one first concept term (6) from medical history information (8) represented by a data record (7) characterized by comprising the report at least partially completed (5).
    [0005]
    5. SYSTEM, according to claim 1, characterized in that the at least one second concept term (11) is associated with a specific section of the report, and the indicator (4) is arranged to indicate which section the at least a second concept term (11) is associated.
    [0006]
    6. SYSTEM, according to claim 1, in which the system is characterized by comprising a user interface element (9) to allow a user to indicate that the report (5) is completed, and in which the indicator (4 ) is arranged to indicate at least one second concept term found (11) in response to a user indication that the report (5) is complete.
    [0007]
    7. SYSTEM, according to claim 1, further characterized by comprising a second concept term classifier (12) to classify at least one second concept term (11) by type of concept, and in which the indicator (4 ) is prepared to indicate the type of concept for a second concept term (11).
    [0008]
    8. SYSTEM, according to claim 1, in which the associations of the association sets (1) have different intensities, and additionally characterized by comprising a second concept term classifier (13) to classify at least one second concept term (11) found by the searcher for missing concept terms (3), based on the intensities of the associations by which at least a second concept term (11) was found.
    [0009]
    9. SYSTEM, according to claim 1, further characterized by comprising an association generator (14) to create an association based on the co-occurrence frequencies of the first and second concept terms in a set of reports.
    [0010]
    10. WORK STATION, characterized by comprising the system as defined in claim 1.
    [0011]
    11. METHOD OF AID TO CREATE REPORTS, characterized by understanding; - the extraction (2 O1) of at least one first term of concept related to a specific individual, from a report at least partially completed related to the specific individual; - the discovery (2 O2) of at least a second concept term associated with at least one first concept term based on a set of associations, where at least a second concept term is absent from the report at least partially completed , in which an association of the set of associations associates at least one first concept term in a knowledge domain with at least one second concept term in the knowledge domain, where the association indicates a frequent co-occurrence of at least one first concept term and at least a second concept term in knowledge domain reports; - the indication (2 O3) of at least one second concept term found to a user, and allow a user to create at least part of the report (5), indicating the found at least one second concept term (11) during a process creation of at least part of the report (5), in which at least a second concept term (11) comprises an anatomical region term, and in which the system additionally comprises an image viewer (15) configured to indicate a view of an individual's anatomical region based on the term anatomical region, where the anatomical region corresponds to the term anatomical region.
    类似技术:
    公开号 | 公开日 | 专利标题
    BR112013002534B1|2020-12-22|report creation aid system, workstation, and report creation aid method
    Moola et al.2015|Conducting systematic reviews of association |: the Joanna Briggs Institute's approach
    RU2620890C2|2017-05-30|Report generation based on image data
    JP6461909B2|2019-01-30|Context-driven overview view of radiation findings
    JP6914839B2|2021-08-04|Report content context generation for radiation reports
    RU2711305C2|2020-01-16|Binding report/image
    CN105940401A|2016-09-14|Context sensitive medical data entry system
    CN106415555B|2019-10-18|Associated system and method for pathologists report and radiological report
    RU2741734C2|2021-01-28|Long-term patient health profile for random findings
    EP2862109A1|2015-04-22|System and method for generating textual report content using macros
    D'Agostino et al.2017|Psychometric Evaluation of the D‐C atch, an Instrument to Measure the Accuracy of Nursing Documentation
    JP2019509572A|2019-04-04|Methods for generating narrative reports from viable clinical pathways
    JP6297571B2|2018-03-20|Support for clinical decisions
    US10318092B2|2019-06-11|Medical records visualization system for displaying related medical records in clusters with marked interrelationships on a time line
    Davies et al.2016|Developing a registry for thyroid incidentalomas: lessons learned and the path forward
    Papež et al.2017|Evaluating OpenEHR for storing computable representations of electronic health record phenotyping algorithms
    Tritter et al.2014|What are tests for? The implications of stuttering steps along the US patient pathway
    US10832805B2|2020-11-10|Method and system for generating electronic medical reports
    CN108984587A|2018-12-11|Information processing unit, information processing method, information processing system and storage medium
    JP5816321B2|2015-11-18|Information processing apparatus, information processing system, information processing method, and program
    CN109416944A|2019-03-01|Use ontological evaluation decision tree
    WO2020241616A1|2020-12-03|Automated diagnosis assisting processing method for orthodontic practice
    US20150066535A1|2015-03-05|System and method for reporting multiple medical procedures
    US20200395113A1|2020-12-17|Implementing and verifying compliance with current standards of care
    JP2021531537A|2021-11-18|Cognitive analysis and disambiguation of electronic medical records for presentation of relevant information about medical plans
    同族专利:
    公开号 | 公开日
    EP2601608A1|2013-06-12|
    EP2601608B1|2019-10-09|
    US20130124527A1|2013-05-16|
    JP2013536503A|2013-09-19|
    WO2012017418A1|2012-02-09|
    JP5982368B2|2016-08-31|
    CN103069423B|2018-07-17|
    CN103069423A|2013-04-24|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US6047259A|1997-12-30|2000-04-04|Medical Management International, Inc.|Interactive method and system for managing physical exams, diagnosis and treatment protocols in a health care practice|
    US6542905B1|1999-03-10|2003-04-01|Ltcq, Inc.|Automated data integrity auditing system|
    US20050033121A1|2001-12-28|2005-02-10|Modrovich Ivan E.|Diagnostic information systems|
    US20030154085A1|2002-02-08|2003-08-14|Onevoice Medical Corporation|Interactive knowledge base system|
    CA2486089C|2002-05-15|2013-12-17|Tommy J. Morris|System and method for handling medical information|
    US20050131738A1|2002-05-15|2005-06-16|Morris Tommy J.|System and method for handling medical information|
    US20040019501A1|2002-07-27|2004-01-29|White Scott B.|Patient scheduling, tracking and status system|
    US20050065813A1|2003-03-11|2005-03-24|Mishelevich David J.|Online medical evaluation system|
    US20050197545A1|2004-03-02|2005-09-08|Hoggle John M.|System and method for disease management|
    US20050273363A1|2004-06-02|2005-12-08|Catalis, Inc.|System and method for management of medical and encounter data|
    US20060004745A1|2004-06-04|2006-01-05|Agfa Corporation|Structured reporting report data manager|
    US20060020466A1|2004-07-26|2006-01-26|Cousineau Leo E|Ontology based medical patient evaluation method for data capture and knowledge representation|
    US20060206358A1|2005-02-28|2006-09-14|Beaver Stephen J|Medical data reporting systems, methods and devices|
    JP2006260318A|2005-03-18|2006-09-28|Hitachi Medical Corp|Diagnostic reading report input support method and system|
    US20070067185A1|2005-09-16|2007-03-22|Halsted Mark J|Medical diagnosis feedback tool|
    US8301461B2|2006-01-30|2012-10-30|Bruce Reiner|Method and apparatus for generating a radiologist quality assurance scorecard|
    US20120296675A1|2006-02-13|2012-11-22|Silverman David G|Method and System for Assessing, Quantifying, Coding & Communicating a Patient's Health and Perioperative Risk|
    WO2008125996A2|2007-04-13|2008-10-23|Koninklijke Philips Electronics, N.V.|Method and system for determining correlation between clinical events|
    US20080275913A1|2007-05-01|2008-11-06|Van Arragon Paul|Dynamic assignment of statements for selected exams using clinical concepts|
    JP5251064B2|2007-10-11|2013-07-31|富士ゼロックス株式会社|Interpretation report creation support device|
    WO2009064850A2|2007-11-13|2009-05-22|Wasson David S|Method and apparatus for indirect medical consultation|
    US8046226B2|2008-01-18|2011-10-25|Cyberpulse, L.L.C.|System and methods for reporting|
    JP2009223853A|2008-03-19|2009-10-01|Fuji Xerox Co Ltd|Medical diagnostic support apparatus|
    EP2120170A1|2008-05-14|2009-11-18|Algotec Systems Ltd.|Distributed integrated image data management system|
    US8554580B2|2008-10-10|2013-10-08|General Electric Company|Automated management of medical data using expert knowledge and applied complexity science for risk assessment and diagnoses|
    US20110004588A1|2009-05-11|2011-01-06|iMedix Inc.|Method for enhancing the performance of a medical search engine based on semantic analysis and user feedback|
    CN101763468A|2010-01-05|2010-06-30|浙江大学|Method for processing clinical auxiliary examination data|JP2013235392A|2012-05-08|2013-11-21|Kddi Corp|Treatment plan creation support system, program and treatment plan creation support method|
    JP6101563B2|2013-05-20|2017-03-22|株式会社日立製作所|Information structuring system|
    CN104182059A|2013-05-23|2014-12-03|华为技术有限公司|Generation method and system of natural language|
    EP3000064A1|2013-06-05|2016-03-30|Nuance Communications, Inc.|Methods and apparatus for providing guidance to medical professionals|
    US11183300B2|2013-06-05|2021-11-23|Nuance Communications, Inc.|Methods and apparatus for providing guidance to medical professionals|
    US9146980B1|2013-06-24|2015-09-29|Google Inc.|Temporal content selection|
    CN103578074A|2013-11-25|2014-02-12|方正国际软件有限公司|Input assisting system and input assisting method for medical history|
    WO2015079354A1|2013-11-26|2015-06-04|Koninklijke Philips N.V.|System and method of determining missing interval change information in radiology reports|
    JP6826039B2|2015-02-05|2021-02-03|コーニンクレッカ フィリップス エヌ ヴェＫｏｎｉｎｋｌｉｊｋｅ Ｐｈｉｌｉｐｓ Ｎ．Ｖ．|Communication system for dynamic checklists to support radiation reporting|
    TR201908751T4|2015-02-25|2019-07-22|Koninklijke Philips Nv|Detection of missing findings for automatic generation of longitudinal findings image.|
    JP2017033518A|2015-07-31|2017-02-09|キヤノン株式会社|Device, method, system and program|
    US20170083665A1|2015-09-23|2017-03-23|Siemens Healthcare Gmbh|Method and System for Radiology Structured Report Creation Based on Patient-Specific Image-Derived Information|
    CN108292527A|2015-11-30|2018-07-17|皇家飞利浦有限公司|The system of clinical discovery wheel disc-exploration clinical concept|
    US20170337329A1|2016-05-18|2017-11-23|Siemens Healthcare Gmbh|Automatic generation of radiology reports from images and automatic rule out of images without findings|
    US10152571B1|2017-05-25|2018-12-11|Enlitic, Inc.|Chest x-ray differential diagnosis system|
    US11145059B2|2018-11-21|2021-10-12|Enlitic, Inc.|Medical scan viewing system with enhanced training and methods for use therewith|
    US10943681B2|2018-11-21|2021-03-09|Enlitic, Inc.|Global multi-label generating system|
    法律状态:
    2018-10-23| B25D| Requested change of name of applicant approved|Owner name: KONINKLIJKE PHILIPS N.V. (NL) |
    2018-11-13| B25G| Requested change of headquarter approved|Owner name: KONINKLIJKE PHILIPS N.V. (NL) |
    2018-12-26| B06F| Objections, documents and/or translations needed after an examination request according art. 34 industrial property law|
    2020-01-28| B06U| Preliminary requirement: requests with searches performed by other patent offices: suspension of the patent application procedure|
    2020-10-06| B09A| Decision: intention to grant|
    2020-10-06| B15K| Others concerning applications: alteration of classification|Free format text: A CLASSIFICACAO ANTERIOR ERA: G06F 19/00 Ipc: G16H 10/00 (2018.01), G16H 15/00 (2018.01), G16Z 9 |
    2020-12-22| B16A| Patent or certificate of addition of invention granted|Free format text: PRAZO DE VALIDADE: 20 (VINTE) ANOS CONTADOS A PARTIR DE 05/08/2011, OBSERVADAS AS CONDICOES LEGAIS. |
    优先权:
    申请号 | 申请日 | 专利标题
    US37083210P| true| 2010-08-05|2010-08-05|
    US61/370,832|2010-08-05|
    PCT/IB2011/053509|WO2012017418A1|2010-08-05|2011-08-05|Report authoring|
    [返回顶部]