Disease treatment competition pattern information retrieval method and device, computer equipment and computer readable storage medium
By constructing a treatment pattern database and using automatic mapping algorithm to match information, the problem of time-consuming collection of competitive pattern information for disease treatment is solved, efficient and accurate information retrieval is achieved, and the needs of pharmaceutical companies and investors are met.
Patent Information
- Application Number
- CN202411130814.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-16
- Publication Date
- 2024-11-19
Smart Images

Figure CN118981484A_ABST
Abstract
Description
Technical Field
[0001] Embodiments of the present invention relate to the field of computer technology, and in particular, to a method, device, computer equipment, and computer-readable storage medium for retrieving information on the competitive landscape of disease treatments. Background Art
[0002] For pharmaceutical companies, in order to explore unmet clinical needs, starting from mechanisms and targets is the basis of research and development, but drug development and preclinical research are difficult, and there are more challenges in the subsequent clinical stage. Based on this, if there is a platform that continuously tracks changes in the treatment landscape of the global new drug market, it can not only help pharmaceutical companies optimize their research and development strategies, but also meet the needs of investors to screen and explore projects, killing two birds with one stone.
[0003] However, there is currently no comprehensive database product on the market that provides this type of data support. If you want to do a good job of analyzing the competitive situation of disease treatment, you can only spend a lot of time and a team to manually collect this type of information. This is not only time-consuming, but also a great test of the professionalism of team members. It can seriously lead to information errors, which in turn affect the company's strategic decisions, such as affecting the fate of a pharmaceutical company's pipeline in the next few years, or even the fate of the company a few years later.
[0004] Therefore, it is an urgent problem to develop a comprehensive information platform that can display the progress of current treatment plans so that users can obtain progress information of relevant trial plans, and then plan the medical strategy for each new drug research and development pipeline, analyze the current status of clinical needs and market competition, etc. Summary of the invention
[0005] The purpose of the present invention is to provide a disease treatment competition landscape information retrieval method, device, computer equipment and computer-readable storage medium to solve the defects of existing biomedical information collection, such as long time consumption and incompleteness, resulting in high cost and coarse granularity of treatment landscape situation information survey and failure to meet demand.
[0006] In a first aspect, the present invention provides a method for retrieving information on the competitive landscape of disease treatments, comprising: Obtain user search terms for querying information on the competitive landscape of disease treatments; Through the pre-built treatment pattern database, target information matching the user's search terms is screened out; the target information includes at least one of the following: medical evidence chain information, R&D progress information, combination information, predicted time to market, and special review information for drugs in segmented populations; Send the target information to the client for display.
[0007] In some embodiments of the present invention, before obtaining user search terms for querying disease treatment competition landscape information, it also includes: obtaining biomedical information, which includes clinical trial information, evidence-based results information, review and approval information, and guideline inclusion information; performing key entity recognition and standardization processing on the biomedical information to obtain key entity information; constructing a population segmentation dictionary based on the key entity information; wherein each population segmentation included in the population segmentation dictionary is composed of at least two entity fields of indication, staging, pathology, biomarkers, and baseline characteristics; based on the population segmentation dictionary, determining the medical evidence chain information, research and development progress information, scheme combination information, predicted time to market, and special review information of the drug under each population segment; and collecting the medical evidence chain information, research and development progress information, scheme combination information, predicted time to market, and special review information of the drug under each population segment to construct a treatment landscape database.
[0008] In some embodiments of the present invention, key entity recognition and standardization processing are performed on biomedical information to obtain key entity information, including: inputting the biomedical information into a trained entity recognition model and outputting initial entity information; standardizing the initial entity information to obtain key entity information; wherein the key entity information includes but is not limited to at least one of the following: indications, staging, pathology, biomarkers, baseline characteristics, treatment thread, generic name of the drug, registration number, approval agency, and approval time; the trained entity recognition model is trained using preset label content, and the preset label content includes: indications, staging, pathology, biomarkers, baseline characteristics, treatment thread, and regimen elements.
[0009] In some embodiments of the present invention, a dictionary of segmented populations is constructed based on key entity information, including: counting key entity information according to entity fields that constitute the segmented populations to obtain statistical results; if the statistical results meet the preset conditions for adding new segmented populations, retrieving and updating the pre-stored segmented population relationship table, the segmented population relationship table being constructed based on the parent-child relationship of each segmented population; if the statistical results do not meet the conditions for adding new populations, retrieving the pre-stored segmented population relationship table; for the currently retrieved segmented population relationship table, using a preset automatic mapping algorithm to automatically match each key entity information to the segmented population relationship table, and constructing a segmented population dictionary.
[0010] In some embodiments of the present invention, based on the dictionary of segmented populations, the medical evidence chain information, research and development progress information, scheme combination information, predicted time to market and special review information of the drug under each segmented population are determined, including: for each segmented population in the dictionary of segmented populations, the treatment thread and the generic name of the drug associated therewith are extracted from the key entity information; the interrelated segmented populations, treatment threads and generic names of the drugs are used as data primary key information; for the data primary key information, the clinical trial information, the evidence-based result information, the review and approval information and the guideline inclusion information are associated with the registration number in the clinical trial information to obtain the medical evidence chain information; For the data primary key information, the highest R&D stage, as well as the milestone time and event status corresponding to the highest R&D stage are calculated to obtain the R&D progress information; for the data primary key information, the type of scheme usage is determined as the scheme combination information; wherein the scheme usage types include single-drug schemes and combination schemes; for the data primary key information, the mean historical time to market or the median historical time to market of similar drugs is calculated as the predicted time to market; for the data primary key information, at least one of the review document information, the information released by the regulatory agency, and the news information is obtained, and when the preset special approval keywords are monitored, a special review tag is generated as the special review information.
[0011] In some embodiments of the present invention, the highest research and development stage, as well as the milestone time and event status corresponding to the highest research and development stage are calculated for the data primary key information to obtain research and development progress information, including: for the data primary key information, based on the medical evidence chain information, determining the research and development stage corresponding to the clinical trial conducting region and / or the review region; wherein the research and development stages from high to low at least include: approval for marketing, application for marketing, Phase III clinical trials, Phase II / Phase III clinical trials, Phase II clinical trials, Phase I / II clinical trials, and Phase I clinical trials; based on the research and development stages corresponding to the clinical trial conducting region and / or the review region, determining the highest research and development stage of the data primary key information in different regions; determining the time to enter the highest research and development stage as the milestone time; determining the announcement status of the highest research and development stage as the event status, and obtaining the research and development progress information through aggregation.
[0012] In some embodiments of the present invention, a scheme attribute label is determined for data primary key information as scheme combination information, including: determining biomedical information associated with the data primary key information; inputting the biomedical information into a trained text classification model, and outputting a target sentence describing a treatment scheme in the biomedical information; inputting the target sentence into a trained pointer model, so that the pointer model extracts a scheme from the target sentence based on preset prompt information, and outputs a treatment scheme corresponding to the prompt information; if the treatment scheme corresponding to the prompt information is a first treatment scheme that exists independently, then determining a single-drug scheme as scheme combination information; if the treatment scheme corresponding to the prompt information is an associated second treatment scheme, then determining a combination scheme as scheme combination information.
[0013] In a second aspect, the present invention provides a disease treatment competition landscape information retrieval device, comprising: A term acquisition module, used to acquire user search terms for querying information on the competitive landscape of disease treatment; An information screening module is used to screen out target information that matches the user's search terms through a pre-built treatment pattern database; wherein the target information includes at least one of the following: medical evidence chain information, research and development progress information, combination information, predicted time to market, and special review information of the drug under segmented populations; The information sending module is used to send the target information to the client for display by the client.
[0014] In a third aspect, the present invention further provides a computer device, comprising: one or more processors; A memory; and one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the processor to implement the above-mentioned disease treatment competition landscape information retrieval method.
[0015] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, which is loaded by a processor to execute the steps in the disease treatment competition landscape information retrieval method.
[0016] In a fifth aspect, an embodiment of the present invention provides a computer program product or a computer program, wherein the computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method provided in the first aspect above.
[0017] The above-mentioned disease treatment competition landscape information retrieval method, device, computer equipment and computer-readable storage medium, the server obtains the user's search terms and calls a pre-built treatment landscape database to screen out medical evidence chain information, research and development progress information, scheme combination information, predicted time to market and / or special review information that match the user's search terms, and then sends the target information required by the user to the client for display, thereby improving the efficiency of obtaining disease treatment competition landscape information, meeting the user's research needs for the treatment landscape situation, and saving data research costs. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.
[0019] Figure 1 A schematic diagram of a scenario of a disease treatment competition landscape information retrieval method in an embodiment of the present invention; Figure 2 is a flow chart of a disease treatment competition landscape information retrieval method in an embodiment of the present invention; Figure 3 is a schematic diagram of the structure of a disease treatment competition landscape information retrieval device in an embodiment of the present invention; Figure 4 It is a schematic diagram of the structure of a computer device in an embodiment of the present invention. DETAILED DESCRIPTION
[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.
[0021] It should be noted that, in the description of this application, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of this application, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.
[0022] First, the disease treatment competition landscape information retrieval method provided by the embodiment of the present invention can be applied to Figure 1In the disease treatment competition pattern information retrieval system shown. Among them, the disease treatment competition pattern information retrieval system includes a client 102 and a server 104. The client 102 can be a device that includes both receiving and transmitting hardware, that is, a device with receiving and transmitting hardware that can perform two-way communication on a two-way communication link. Such a device may include: a cellular or other communication device, a cellular or other communication device with a single-line display or a multi-line display. The client 102 can specifically be a desktop terminal or a mobile terminal, and the client 102 can specifically be one of a mobile phone, a tablet computer, and a laptop computer. The server 104 can be an independent server, or a server network or server cluster composed of servers, which includes but is not limited to a computer, a network host, a single network server, a plurality of network server sets, or a cloud server composed of a plurality of servers. Among them, the cloud server is composed of a large number of computers or network servers based on cloud computing (Cloud Computing). In addition, a communication connection is established between the client 102 and the server 104 through a network, and the network can specifically be any one of a wide area network, a local area network, and a metropolitan area network.
[0023] Secondly, those skilled in the art will understand that Figure 1 The application environment shown in the figure is only one application scenario applicable to the present application solution and does not constitute a limitation on the application scenario of the present application solution. Other application environments may also include Figure 1 More or fewer devices are shown in . For example, Figure 1 Only one server is shown. It is understandable that the disease treatment competition landscape information retrieval system may also include one or more other devices, which are not specifically limited here. In addition, the disease treatment competition landscape information retrieval system may also include a memory for storing data, such as storing at least one of medical evidence chain information, R&D progress information, scheme combination information, predicted time to market, and special review information.
[0024] certainly, Figure 1 The scenario diagram of the disease treatment competition landscape information retrieval system shown is merely an example. The disease treatment competition landscape information retrieval system and scenario described in the embodiment of the present invention are intended to more clearly illustrate the technical solution of the embodiment of the present invention, and do not constitute a limitation on the technical solution provided by the embodiment of the present invention. A person of ordinary skill in the art can appreciate that with the evolution of the disease treatment competition landscape information retrieval system and the emergence of new business scenarios, the technical solution provided by the embodiment of the present invention is equally applicable to similar technical problems.
[0025] Here, before describing in detail the technical solution provided by the embodiment of the present application, it is necessary to understand that with the emergence of the wave of innovative drugs in China in the past five years, most of China's original new drugs have entered the clinical trial stage one after another. However, once entering the clinical stage, pharmaceutical companies generally face a series of challenges such as conceptual verification, exploration of optimal dosing, efficacy / safety, etc. In this process, enterprises often face various uncertainties both inside and outside. For example, within the enterprise, due to the further understanding of the "product-population" characteristics, and in the face of future commercialization needs and verification of clinical value, the indication positioning and target population of the product will be continuously optimized and adjusted. In the external environment, the entire treatment pattern is also constantly changing over time, and new drugs and target populations continue to impact the existing pattern. Therefore, whether from the perspective of pharmaceutical research and development, BD, strategy or investment and financing, it is very necessary to continuously track and deeply study the changes in the treatment pattern of the global new drug market.
[0026] In order to achieve the above goals, that is, to understand "one's own" products while also being able to clearly understand competing products, the research information must include not only key clinical trial information and clinical trial results information, but also corresponding approval information, key clinical information supporting approval, special channel information, and inclusion in core guidelines. However, without the support of a comprehensive database product, manual research and collection of this information takes a long time, ranging from a few weeks to several months. The difficulty is not only reflected in the time consumption, but also in the certain threshold for accurate processing of medical information. Specifically, differences in understanding of details such as clinical trials, literature, guidelines, and approval information by people with different backgrounds and experiences may lead to bias in the final information.
[0027] On the other hand, looking at the existing drug-related database products, the classification of diseases is mostly refined to common "disease categories", and the characteristics of the subdivided population are not clearly shown, especially the treatment threads and biomarkers. For example, for the treatment pattern of the popular disease "non-small cell lung cancer", its track is already very crowded. By querying "non-small cell lung cancer", there will be a feeling of "redundancy". If you want to further divide and refine it to "first-line treatment of exon 20 insertion", it will require huge labor costs, and maintaining updates is also a major difficulty. From the perspective of investment and financing, the logic is the same. In other words, using the existing drug-related database products for target queries, the obtained "competitive products" or "high-value products" will inevitably have more impurities. If you want to do a good job of analyzing the competitive situation of disease treatment, you still need to spend a lot of team and time to manually collect more information.
[0028] Furthermore, due to the large amount of such information, the information sources are mostly in unstructured formats such as text, web pages, and documents. The collected information cannot be used directly, and it needs to rely on experienced personnel to accurately interpret it, and then spend a lot of time to manually structure it, and finally organize the information into standardized data according to certain standards. Moreover, the collection, processing, and association of information from all parties are extremely difficult and professional. In addition, this type of information may change every day, and it needs to be updated at a high frequency to meet the needs of competition monitoring and early warning. This also leads to the fact that there is still no comprehensive product with a similar concept on the market.
[0029] Based on this, the present application embodiment proposes a disease treatment competition landscape information retrieval method, which can be configured in a computer device to integrate a comprehensive information platform that can display the progress of the current treatment plan. Figure 2 , is a flow chart of a disease treatment competition landscape information retrieval method provided by an embodiment of the present invention. This embodiment mainly applies the method to the above Figure 1 Taking the server 104 in the example, the method includes steps S201 to S203, which are as follows: S201, obtaining user search terms for querying information on the competitive landscape of disease treatments.
[0030] Among them, the user search terms can be related words used to summarize the competitive landscape information of the disease treatment to be retrieved, including but not limited to: indication (such as chronic lymphocytic leukemia), clinical stage (such as early, locally advanced), pathology (such as squamous cell carcinoma, adenocarcinoma), biomarkers (such as HR positive, BRCA mutation), baseline characteristics (such as age, gender), treatment thread (such as first-line treatment, neoadjuvant therapy), drugs (such as toripalimab), targets (such as PD1), pharmacological types (such as antibody-drug conjugates, mRNA vaccines), drug categories (such as compounds, peptides), highest research and development stage (such as approval for marketing, Phase III clinical trials), and regimen labels (such as monotherapy, combination therapy).
[0031] In a specific implementation, a user may submit one or more user search terms for querying information on the competitive landscape of disease treatments on the client 102, so that after receiving the search terms submitted by the user, the client 102 sends the user search terms to the server 104, instructing the server 104 to retrieve and feedback target information matching the user search terms based on the user search terms.
[0032] In one embodiment, before step S201, it also includes: obtaining biomedical information, which includes clinical trial information, evidence-based results information, review and approval information, and guideline inclusion information; performing key entity recognition and standardization processing on the biomedical information to obtain key entity information; constructing a population segmentation dictionary based on the key entity information; wherein each population segmentation included in the population segmentation dictionary is composed of at least two entity fields of indication, stage, pathology, biomarker, baseline characteristics, etc.; based on the population segmentation dictionary, determining the medical evidence chain information, research and development progress information, scheme combination information, predicted time to market, and special review information of the drug under each population segment; and collecting the medical evidence chain information, research and development progress information, scheme combination information, predicted time to market, and special review information of the drug under each population segment to construct a treatment pattern database.
[0033] Among them, the purpose of obtaining clinical trial information is to clarify which key clinical trials have been or are ongoing. The purpose of obtaining evidence-based results information is to clarify the efficacy and safety data related to clinical trials. The purpose of obtaining review and approval information is to clarify which treatment options have been approved for marketing in various regions. The purpose of obtaining guideline inclusion information is to clarify whether the corresponding treatment option has been included in the important guidelines and how it has been included. The purpose of obtaining these four types of information is not only to generate medical evidence chain information so that users can retrieve the full life cycle content of clinical trial projects, such as users can query all approved drugs and medical evidence chains under a certain indication, including guideline inclusion, key clinical trials, and test results; but also to integrate the information content generated by the above stages. Compared with analyzing the content of a certain stage alone, the results are undoubtedly more comprehensive and more accurate, and can better meet users' high-precision retrieval needs for information on the competitive landscape of disease treatment.
[0034] In a specific implementation, before constructing the treatment pattern database, the server 104 can obtain various types of biomedical information through compliant methods, such as compliant access to specific databases and websites. For example, clinical trial information can be obtained through a clinical trial registration platform (such as a drug clinical trial registration and information disclosure platform), evidence-based results information can be obtained through a medical literature database or literature citation aggregation site (such as a Pubmed database, major international conferences, news announcements), guideline inclusion information can be obtained through guideline collection platforms in various countries, and review and approval information can be obtained through the relevant official websites and news announcements of drug regulatory agencies in various countries.
[0035] Further, after the server 104 obtains the biomedical information, it may pre-process the biomedical information, including but not limited to: noise removal, data cleaning, unified format, data normalization, etc. Among them, noise removal refers to removing useless or erroneous information in the data, such as special characters, HTML tags, repeated content, etc., to reduce interference and improve the consistency and accuracy of the data. Data cleaning refers to cleaning the data, including removing blank characters, missing value processing, correcting erroneous data, etc., to improve the quality of the data and reduce ambiguity. Unified format refers to unifying data from different sources and different representations into a consistent format, such as unifying dates into the same date format, unifying units into the same measurement unit, etc., to eliminate ambiguity caused by different forms of representation. Data normalization refers to normalizing numerical data, scaling it to a unified range, reducing the scale difference between different data, and avoiding adverse effects on subsequent processing.
[0036] Furthermore, after the server 104 completes the preprocessing operation, it can perform entity recognition on the preprocessed biomedical information. Here, entity recognition can be assisted by a pre-trained entity recognition model, and the data required for training may include various dictionaries pre-constructed in this application, including but not limited to: indication dictionary, staging dictionary, biomarker dictionary, pathology dictionary, baseline dictionary, treatment thread dictionary, and treatment plan dictionary. The various dictionaries mentioned here are all constructed based on preset label content: indication, staging, pathology, biomarker, baseline characteristics, treatment thread, and plan elements, which will be described in detail in subsequent embodiments.
[0037] Furthermore, after the server 104 performs entity recognition, the entity recognition results obtained need to be converted into a standard form for subsequent analysis and processing. Here, the standardization processing methods include but are not limited to: case standardization, spelling correction, abbreviation expansion, date and time format standardization, and digital format standardization. The role of standardization processing is to improve the consistency and comparability of data, thereby improving the accuracy of subsequent analysis, making the treatment pattern database constructed by the embodiment of the present application more accurate, thereby satisfying the user's high-precision retrieval needs for disease treatment competition pattern information.
[0038] Furthermore, the server 104 standardizes the entity recognition results, and after obtaining the key entity information, it can construct a segmented population dictionary based on the key entity information, and then use the segmented population dictionary to determine the medical evidence chain information, research and development progress information, scheme combination information, predicted time to market, and special review information of the drug under each segmented population required for constructing the treatment pattern database. Here, the steps of determining the segmented population dictionary and the steps of determining various types of information involved will be described in detail below with the help of embodiments.
[0039] In one embodiment, key entity recognition and standardization are performed on biomedical information to obtain key entity information, including: inputting the biomedical information into a trained entity recognition model and outputting initial entity information; standardizing the initial entity information to obtain key entity information; wherein the key entity information includes but is not limited to at least one of the following: indication, staging, pathology, biomarker, baseline characteristics, treatment thread, generic name of the drug, registration number, approval agency, and approval time; the trained entity recognition model is trained using preset label content, and the preset label content includes: indication, staging, pathology, biomarker, baseline characteristics, treatment thread, and regimen elements.
[0040] In the specific implementation, in order to obtain key entity information, the indication dictionary, staging dictionary, biomarker dictionary, pathology dictionary, baseline dictionary, treatment thread dictionary and treatment plan dictionary built based on the preset label content can be used to achieve efficient labeling of model training samples, and then the model training samples are used to complete the training operation of the entity recognition model to obtain the trained entity recognition model. Here, the preset label content is used to train the entity recognition model, and its role is to ensure the uniformity and centralization of entity labeling, thereby improving the accuracy of entity recognition results.
[0041] Further, after the server 104 has trained and obtained a callable entity recognition model, the current biomedical information to be analyzed can be input into the trained entity recognition model so that the model outputs the initial entity information in the biomedical information, and then the currently obtained initial entity information is standardized to obtain the key entity information. Here, the "indications" contained in the key entity information mean the reasons why the patient is in an unhealthy state, including: breast cancer, endometrial cancer, etc. "Staging" means the stage or phase of the disease, including: early, locally advanced, advanced, etc. "Pathology" means the changes and characteristics of the disease in cells, tissues and organ structures, including: squamous cell carcinoma, adenocarcinoma, etc. "Biomarkers" mean the changes and characteristics of the disease in proteins, genes, biochemical indicators, etc., including: HR positivity, BRCA mutations, etc. "Baseline characteristics" means describing the disease characteristics that cannot be classified as staging, pathology, or biomarkers when the patient is clinically enrolled, including: demographic characteristics (age, gender, region), physical status, previous treatment, etc. "Treatment thread" means the order in which patients receive treatment, including: neoadjuvant therapy, adjuvant therapy, first-line treatment, second-line treatment, etc. "Generic name of drug" means the non-proprietary name of the drug, which is used to identify the main active ingredient of the drug, including: toripalimab, adagracib, etc. "Registration number" is a unique identification number assigned by the drug regulatory authority of a country or region after the drug is approved in a country or region, including: NCTxxxxxxxx. "Approval agency" refers to the national or regional drug regulatory authority responsible for reviewing and approving drugs for marketing, such as the U.S. Food and Drug Administration (FDA), the European Medicines Agency (EMA), the China National Medical Products Administration (NMPA), etc. "Approval time" refers to the specific date on which the drug is approved for marketing by the regulatory authority.
[0042] It should be noted that the entity recognition model in the embodiment of the present application can be constructed using an "encoder-decoder" structure. The indications, staging, pathology, biomarkers, baseline characteristics, and treatment threads contained in the preset label content have the same meaning as the indications, staging, pathology, biomarkers, baseline characteristics, and treatment threads in the key entity information. In addition, the program elements refer to the elements that make up the treatment plan, including single drugs (such as olaparib), program abbreviations (such as R-CHOP), major drug categories (such as BTK inhibitors), non-drug treatments (such as radiation therapy), etc.
[0043] In one embodiment, a dictionary of segmented populations is constructed based on key entity information, including: counting key entity information according to entity fields constituting the segmented populations to obtain statistical results; if the statistical results meet the preset conditions for adding new segmented populations, retrieving and updating the pre-stored segmented population relationship table, the segmented population relationship table being constructed based on the parent-child relationship of each segmented population; if the statistical results do not meet the conditions for adding new populations, retrieving the pre-stored segmented population relationship table; for the currently retrieved segmented population relationship table, using a preset automatic mapping algorithm to automatically match each key entity information to the segmented population relationship table, and constructing a segmented population dictionary.
[0044] Among them, the dictionary of subdivided populations is constructed based on a preset automatic mapping algorithm, and the mapping logic can adopt a parent-child relationship, inclusion / exclusion of biomarkers, etc. Its construction function is to save the workload of manually binding data, and at the same time, it is conducive to adjusting the population category at any time according to the trend changes of the subdivided population, without the need to re-label the original data. In other words, the entity fields selected to form the subdivided population can be set according to actual needs. For example, the subdivided population involved in the embodiment of the present application is currently composed of three entity fields of "indication-staging-biomarker", but those skilled in the art can use any two, three or even four entity fields of "indication", "staging", "pathology", "biomarker", and "baseline characteristics", such as "indication-staging-biomarker-baseline characteristics", or "indication-staging-pathology", etc., which are not limited to the specific embodiments of the present application.
[0045] In the specific implementation, in order to build a dictionary for segmented populations, a segmented population relationship table must be built first. The relevant logic for building the segmented population relationship table includes six categories: indications, biomarkers, staging, pathology, baseline, and treatment thread. Each category includes three fields: "name", "subfield", and "whether to automatically update subfields". A total of 18 fields can be edited. In addition, the relationship between fields of different categories is "and", and the relationship within the same category is "or". A segmented population can have multiple groups of segmented population generation logic.
[0046] For example, "PD-L1 positive advanced non-small cell lung cancer" includes: indications (name: non-small cell lung cancer; subfields: squamous non-small cell lung cancer, non-squamous non-small cell lung cancer, etc.; whether to automatically update subfields: yes), staging (name: advanced stage, stage IV; subfields: stage IVa, stage IVb; whether to automatically update subfields: yes), biomarkers (name: PD-L1 positive; subfields: PD-L1 expression ≥ 50%, etc.; whether to update subfields: yes). This means that for data marked as "non-small cell lung cancer, and its subfields (indications)" and "advanced or stage IV, and its subfields (staging)" and "PD-L1 positive, and its subfields (biomarkers)", the subdivided population "PD-L1 positive advanced non-small cell lung cancer" needs to be generated, and the relevant fields are automatically updated when the subfields are modified in their corresponding fields.
[0047] Finally, the relationship table of the subdivided populations is presented as follows: "Non-small cell lung cancer" is expanded to include (1) "early non-small cell lung cancer" and (2) "advanced non-small cell lung cancer", etc., (1) "early non-small cell lung cancer" is expanded to include "EGFR mutation early non-small cell lung cancer", "KRAS mutation early non-small cell lung cancer", "PD-L1 positive early non-small cell lung cancer", etc., (2) "advanced non-small cell lung cancer" is expanded to include "EGFR mutation advanced non-small cell lung cancer", "KRAS locally advanced non-small cell lung cancer", "PD-L1 positive advanced non-small cell lung cancer", etc.
[0048] Furthermore, the segmented population relationship table constructed according to actual needs can be stored in the server 104 for use at any time, that is, when the server 104 executes the segmented population dictionary construction command or executes the segmented population dictionary update command, it can first count the key entity information currently collected and identified according to the entity fields that constitute the segmented population, and then determine whether to update the pre-stored segmented population relationship table based on the statistical results. If it needs to be updated, it can be retrieved and generated. If it does not need to be updated, the historically stored segmented population relationship table is directly retrieved. Finally, the preset automatic mapping algorithm is used to automatically match each key entity information to the segmented population relationship table, and the segmented population dictionary can be obtained.
[0049] Exemplarily, server 104 determines through statistics that the number of "locally late stages" under the key entity information "stage" exceeds a preset threshold, and the entity field granularity of the historically stored segmented population relationship table has not reached "locally late stage", then it can be determined that the current statistical result meets the preset conditions for adding new segmented populations, and the segmented population relationship table can be updated, and its "stage" can be further refined to "locally late stage", and corresponding segmented populations that conform to the composition pattern can be generated based on the identification of other key entity information, so as to obtain a new segmented population relationship table.
[0050] For example, in combination with the above description, it can be seen that for data where the key entity information obtained by identification includes "non-small cell lung cancer + PD-L1 positive + advanced stage", the automatically generated subdivided population is: "PD-L1 positive advanced non-small cell lung cancer", and for data where the key entity information obtained by identification includes "non-squamous non-small cell lung cancer + PD-L1 expression ≥50% + advanced stage", the subdivided population is also generated as: "PD-L1 positive advanced non-small cell lung cancer".
[0051] Therefore, the server 104 adopts a preset automatic mapping algorithm to automatically match the key entity information to the segmented population relationship table, so that each segmented population in the segmented population relationship table can be associated with the corresponding biomedical information, achieving the purpose of information association and traceability, thereby saving the workload of manual binding of data, and realizing automatic classification and aggregation of segmented populations.
[0052] In one embodiment, based on the dictionary of segmented populations, the medical evidence chain information, research and development progress information, scheme combination information, predicted time to market and special review information of the drug under each segmented population are determined, including: for each segmented population in the dictionary of segmented populations, the treatment thread and the generic name of the drug associated with it are extracted from the key entity information; the interrelated segmented populations, treatment threads and generic names of the drugs are used as data primary key information; for the data primary key information, the clinical trial information, evidence-based result information, review and approval information and guideline inclusion information are associated according to the registration number in the clinical trial information to obtain the medical evidence chain information; for each segmented population in the dictionary of segmented populations, the treatment thread and the generic name of the drug are extracted from the key entity information; the interrelated segmented populations, treatment threads and generic names of the drugs are used as data primary key information; for each segmented population in the dictionary of segmented populations, the clinical trial information, evidence-based result information, review and approval information and guideline inclusion information are associated according to the registration number in the clinical trial information to obtain the medical evidence chain information; For the data primary key information, calculate the highest R&D stage, as well as the milestone time and event status corresponding to the highest R&D stage, to obtain the R&D progress information; for the data primary key information, determine the type of scheme usage as the scheme combination information; wherein, the scheme usage types include single-drug schemes and combination schemes; for the data primary key information, calculate the historical mean or median time to market of similar drugs as the predicted time to market; for the data primary key information, obtain at least one of the review document information, regulatory agency release information, and news information, and generate a special review tag as special review information when monitoring the presence of preset special approval keywords.
[0053] Among them, the embodiment of the present application proposes that a piece of data can be determined by "segmented population", "treatment thread" and "generic name of drug" to perform data cleaning.
[0054] In the specific implementation, in order to obtain various types of information about drugs in various segmented populations, the server 104 must first determine the data primary key information, and the embodiment of the present application proposes that "segmented population, treatment thread, and generic name of drug" can be defined as the data primary key information, that is, the "treatment thread" is used to further refine the treatment pattern of drugs in the "segmented population", so as to refine the granularity of the disease treatment competition pattern information, thereby meeting the user's high-precision retrieval needs for treatment pattern status information.
[0055] Furthermore, since each segmented population is associated with the key entity information of various biomedical information when constructing the segmented population dictionary in the previous step, when analyzing and obtaining various types of information on drugs under each segmented population, server 104 can include clinical trial information, evidence-based result information, review and approval information, and guidelines into the information according to the key entity information corresponding to the data primary key information - "registration number" to complete the information association and obtain the medical evidence chain information.
[0056] Of course, if the above four types of information are only associated by the server 104 according to the "registration number", there may be a part of the information that cannot be associated because it is not bound to the "registration number", resulting in information omissions. Therefore, the embodiment of the present application proposes that the association results of the server 104 can be manually verified again, such as referring to the test details and other contents for manual judgment, especially the review and approval information. Since the review events are coherent, the official data or news announcements often mention the relevant review events when their status is updated or modified. In addition, the relevant review events are related in terms of segmented populations, drugs and supporting evidence for approval. Therefore, data personnel can consider various factors comprehensively to associate the review data, improve the integrity of the medical evidence chain information, and realize the analysis and association of the medical evidence chain information, which can help enterprise users complete a one-stop indication competition landscape survey and comprehensively and deeply evaluate the clinical value of the pipeline.
[0057] Furthermore, the medical evidence chain information obtained by the server 104 can be used as a basis for analyzing and obtaining R&D progress information, scheme combination information, predicted time to market, and special review information, that is, using its information integrity to improve the accuracy and reliability of subsequent information. In this embodiment, the steps for obtaining R&D progress information and scheme combination information will be described in detail below.
[0058] However, it should be noted that when server 104 analyzes and obtains the predicted time to market, the limited scope of similar drugs may depend on the "segmented population" and "generic name of the drug", and may further depend on at least one of the "drug type", "review agency", "whether it is published", "whether it is the first application", and "special channel", which is not limited in the specific embodiments of this application. In addition, when server 104 analyzes and obtains special review information, the preset special approval keywords include but are not limited to at least one of the following: priority review, accelerated review, fast track, breakthrough therapy, rare disease drug, and emergency use authorization.
[0059] In one embodiment, the highest research and development stage, as well as the milestone time and event status corresponding to the highest research and development stage are calculated for the data primary key information to obtain research and development progress information, including: for the data primary key information, based on the medical evidence chain information, determining the research and development stage corresponding to the clinical trial conducting region and / or the review region; wherein the research and development stages from high to low at least include: approval for marketing, application for marketing, Phase III clinical trial, Phase II / Phase III clinical trial, Phase II clinical trial, Phase I / Phase II clinical trial, Phase I clinical trial; based on the research and development stage corresponding to the clinical trial conducting region and / or the review region, determining the highest research and development stage of the data primary key information in different regions; determining the time to enter the highest research and development stage as the milestone time; determining the announcement status of the highest research and development stage as the event status, and obtaining the research and development progress information through aggregation.
[0060] In the specific implementation, when analyzing and obtaining the R&D progress information, the server 104 can first extract the review and approval information (if any) and clinical trial information under each data primary key information, and then determine the current R&D stage in different regions (such as the world, China, the United States, etc.) based on the review region and / or clinical trial conduction region contained in the information, and then determine the highest R&D stage in different regions according to the value priority (i.e., the order of the R&D stage). Thus, further determining the milestone time and event status can help enterprise users investigate and monitor the R&D progress of similar products under different population segments from the perspective of pipelines, and can also help enterprise users investigate and monitor the R&D progress of pipelines under the same population segment dimension from the perspective of population segments, and can also help enterprise users obtain the milestone time and event status of the corresponding pipeline under a specific population segment when investigating the pipeline, so as to facilitate tracking and evaluation of competitive product information.
[0061] For example, the R&D stages of a “segmented population + drug + treatment thread” in China include: approval for marketing, 1 Phase III study, and 2 Phase II studies, then its “highest R&D stage” in China is “approval for marketing.”
[0062] It should be noted that the event status corresponding to "approval for marketing" includes at least one of the following: conditional approval, regular approval, transfer to regular approval, and revocation; the event status corresponding to "clinical trial stage" includes at least one of the following: recruiting, completed, and suspended; "clinical trial stage" includes but is not limited to: "Phase III clinical trial, Phase II / Phase III clinical trial, Phase II clinical trial, Phase I / II clinical trial, Phase I clinical trial".
[0063] In one embodiment, for data primary key information, a scheme attribute label is determined as scheme combination information, including: determining biomedical information associated with the data primary key information; inputting the biomedical information into a trained text classification model, and outputting a target sentence describing a treatment plan in the biomedical information; inputting the target sentence into a trained pointer model, so that the pointer model extracts a scheme from the target sentence based on preset prompt information, and outputs a treatment plan corresponding to the prompt information; if the treatment plan corresponding to the prompt information is a separate first treatment plan, then determining a single-drug plan as the scheme combination information; if the treatment plan corresponding to the prompt information is an associated second treatment plan, then determining a combination plan as the scheme combination information.
[0064] Among them, the first treatment plan refers to the treatment plan mentioned alone in the target sentence; the second treatment plan refers to the treatment plan that can be used in combination with other treatment plans mentioned in the target sentence.
[0065] The preset prompt information may be preset prompt information related to the treatment plan, including first prompt information and second prompt information. The first prompt information is used to prompt the output of a separate treatment plan, for example, "What is the treatment plan mentioned separately in the sentence?" The second prompt information is used to prompt the output of an associated treatment plan, for example, "What is the common treatment plan combined with treatment plan B?"
[0066] In the specific implementation, the text classification model selected in this embodiment includes but is not limited to any of the following: Support Vector Machines (SVM), Logistic Regression, Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM).
[0067] Furthermore, after the server 104 obtains the biomedical information associated with the data primary key information, it can call the trained text classification model to perform machine analysis on the currently obtained biomedical information, so that the model can perform multi-label classification and annotation on the biomedical information based on the functions learned in the previous training, and then output a sentence describing the treatment plan as the target sentence.
[0068] However, since the output target sentence may contain multiple treatment plans, the multiple treatment plans may have the same parts (generally referring to time, number of people, drug usage and dosage, etc.), or may have different parts (generally drug regimen, usage and dosage, etc.), it is impossible to clearly know whether the multiple treatment plans are used in combination or only separately. Therefore, in order to further clarify the relationship between multiple treatment plans and to accurately extract each treatment plan, further analysis of the target sentence is required.
[0069] Furthermore, the server 104 may input the target sentence into a trained pointer model for machine analysis, so that the model can extract the first treatment plan that exists alone and / or the second treatment plan that exists in association therein under a supervised task setting, and finally count and clean all the treatment plans in the biomedical information.
[0070] Specifically, before the server 104 calls the trained pointer model, it needs to complete the model training. In order to make full use of the advantage of the large amount of original data, the original unlabeled data and some expert-level labeled data can be combined to train a generative model based on a certain type of network architecture. After the model is trained, some pseudo-labeled data can be generated. Then, a small amount of expert-level labeled data and a large amount of generated pseudo-labeled data are combined for supervised model learning. The model structure uses a prompt-based pointer network, inputs prompt information + original sentences, outputs the results of the specified task, and iterates and updates repeatedly to obtain the trained pointer model.
[0071] It should be noted that the prompt-based pointer network is a model architecture for generative text tasks, which is mainly used to handle text generation tasks, such as generating answers to questions, article summaries, etc. In the prompt-based pointer network, the input consists of two parts: one is the question or prompt (prompt), and the other is the target text to be generated. The goal of this model is to generate text content related to the given question or prompt by using it. The model usually uses an encoder-decoder architecture. The encoder encodes the input question or prompt and converts it into a fixed-length vector representing semantic information. The decoder then receives this vector and uses it as context information to start generating the target text. The decoder can use methods such as recurrent neural networks (RNNs) or attention mechanisms for text generation. The key innovation is to use the question or prompt as part of the input. By providing additional question or prompt information, the model can better understand and capture the context and semantics of the target text. This approach helps to alleviate ambiguity and confusion in the generation model and can more accurately output text results that meet expectations. The prompt-based pointer network has great potential in natural language processing tasks, especially in generative text tasks. It can generate various related texts according to different questions or prompts, improving the accuracy and coherence of automatically generated texts.
[0072] Furthermore, after the pointer model has the calling conditions, the server 104 can input the target statement into the trained pointer model so that the model can extract the solution for the target statement based on the prompt information and output the treatment plan corresponding to the prompt information. Specifically, the treatment plan corresponding to the first prompt information output can be used as the first treatment plan, and the treatment plan corresponding to the second prompt information output can be used as the second treatment plan. In this way, the server 104 only needs to determine whether the model output result is the first treatment plan or the second treatment plan to generate the corresponding solution combination information. Thus, obtaining the solution combination information can help enterprise users view the layout of the pipeline combination plan under the segmented population and monitor new opportunities for drug combination.
[0073] S202, through the pre-constructed treatment pattern database, screen out target information that matches the user's search terms; wherein the target information includes at least one of the medical evidence chain information, research and development progress information, combination information, predicted time to market, and special review information of the drug in the segmented population.
[0074] In a specific implementation, the above text describes in detail how to construct a treatment pattern database, so after receiving the user search term, the server 104 can respond to the instruction of the client 102 and filter out target information matching the user search term from the treatment pattern database.
[0075] S203, sending the target information to the client for display by the client.
[0076] In a specific implementation, after the server 104 obtains the target information required by the user through query, the target information can be sent to the client 102, so that the client 102 can display the target information to the user through its display screen to meet the user's search needs. It can be understood that the display form of the target information on the client 102 includes but is not limited to: list, grid view, summary card, map view, chart or graph, etc. In addition, the display order of different types of target information can be set according to actual application requirements, and the specific embodiment of the present application is not limited thereto.
[0077] In the disease treatment competition landscape information retrieval method in the above embodiment, the server obtains the user's search terms and calls a pre-built treatment landscape database to screen out medical evidence chain information, research and development progress information, scheme combination information, predicted time to market and / or special review information that match the user's search terms, and then sends the target information required by the user to the client for display, thereby improving the efficiency of obtaining disease treatment competition landscape information, meeting the user's research needs for the treatment landscape situation, and saving data research costs.
[0078] It should be understood that although Figure 2 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 2 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.
[0079] In order to better implement the disease treatment competition landscape information retrieval method provided in the embodiment of the present invention, based on the disease treatment competition landscape information retrieval method proposed in the embodiment of the present invention, the embodiment of the present invention also provides a disease treatment competition landscape information retrieval device, such as Figure 3 As shown, the disease treatment competition landscape information retrieval device 300 includes: The term acquisition module 310 is used to acquire user search terms for querying disease treatment competition landscape information; The information screening module 320 is used to screen out target information matching the user's search terms through a pre-built treatment pattern database; wherein the target information includes at least one of the following: medical evidence chain information, research and development progress information, combination information, predicted time to market, and special review information of the drug under segmented populations; The information sending module 330 is used to send the target information to the client for display by the client.
[0080] In one embodiment, the disease treatment competition landscape information retrieval device 300 also includes a database construction module for obtaining biomedical information, which includes clinical trial information, evidence-based results information, review and approval information, and guideline inclusion information; performing key entity recognition and standardization processing on the biomedical information to obtain key entity information; constructing a population segmentation dictionary based on the key entity information; wherein each population segmentation included in the population segmentation dictionary is composed of at least two entity fields of indication, stage, pathology, biomarker, baseline characteristics, etc.; based on the population segmentation dictionary, determining the medical evidence chain information, research and development progress information, scheme combination information, predicted time to market, and special review information of the drug under each population segment; and collecting the medical evidence chain information, research and development progress information, scheme combination information, predicted time to market, and special review information of the drug under each population segment to construct a treatment landscape database.
[0081] In one embodiment, the database construction module is also used to input biomedical information into a trained entity recognition model and output initial entity information; the initial entity information is standardized to obtain key entity information; wherein the key entity information includes but is not limited to at least one of the following: indications, staging, pathology, biomarkers, baseline characteristics, treatment thread, generic name of the drug, registration number, approval agency, and approval time; the trained entity recognition model is trained using preset label content, and the preset label content includes: indications, staging, pathology, biomarkers, baseline characteristics, treatment thread, and regimen elements.
[0082] In one embodiment, the database construction module is also used to count key entity information based on the entity fields that make up the segmented population to obtain statistical results; if the statistical results meet the preset conditions for adding new segmented populations, the pre-stored segmented population relationship table is retrieved and updated, and the segmented population relationship table is constructed based on the parent-child relationship of each segmented population; if the statistical results do not meet the conditions for adding new populations, the pre-stored segmented population relationship table is retrieved; for the currently retrieved segmented population relationship table, a preset automatic mapping algorithm is used to automatically match each key entity information to the segmented population relationship table to construct a segmented population dictionary.
[0083] In one embodiment, the database construction module is also used to extract the treatment threads and generic drug names associated with each segmented population in the segmented population dictionary from the key entity information; use the mutually associated segmented populations, treatment threads and generic drug names as data primary key information; for the data primary key information, according to the registration number in the clinical trial information, the clinical trial information, evidence-based results information, review and approval information and guideline inclusion information are associated to obtain medical evidence chain information; for the data primary key information, the highest research and development stage, as well as the milestone time and event status corresponding to the highest research and development stage are calculated to obtain research and development progress information; for the data primary key information, the type of scheme use is determined as scheme combination information; wherein the type of scheme use includes single-drug scheme and combination scheme; for the data primary key information, the mean or median of the historical time to market of similar drugs is calculated as the predicted time to market; for the data primary key information, at least one of the review document information, the information released by the regulatory agency and the news information is obtained, and when the preset special approval keywords are detected, a special review tag is generated as the special review information.
[0084] In one embodiment, the database construction module is also used to determine the research and development stage corresponding to the clinical trial conducting region and / or the review region based on the medical evidence chain information for the data primary key information; wherein the research and development stages from high to low include at least: approval for marketing, application for marketing, phase III clinical trial, phase II / phase III clinical trial, phase II clinical trial, phase I / phase II clinical trial, and phase I clinical trial; based on the research and development stage corresponding to the clinical trial conducting region and / or the review region, determine the highest research and development stage of the data primary key information in different regions; determine the time of entering the highest research and development stage as a milestone time; determine the announcement status of the highest research and development stage as an event status, and collect the research and development progress information.
[0085] In one embodiment, the database construction module is also used to determine the biomedical information associated with the data primary key information; input the biomedical information into a trained text classification model, and output a target sentence in the biomedical information that describes a treatment plan; input the target sentence into a trained pointer model, so that the pointer model extracts the plan from the target sentence based on preset prompt information, and outputs a treatment plan corresponding to the prompt information; if the treatment plan corresponding to the prompt information is a separate first treatment plan, then a single-drug plan is determined as the combination plan information; if the treatment plan corresponding to the prompt information is an associated second treatment plan, then a combination plan is determined as the combination plan information.
[0086] The above embodiments illustrate that the present application builds a comprehensive information platform that can display the progress of current treatment plans, so that users can obtain progress information of relevant test plans. This can not only solve the defects of the existing biomedical information collection being time-consuming and incomplete, resulting in high cost and coarse granularity of treatment pattern situation information research and inability to meet the needs, but also can generate certain value for the development of the industry, promote the rapid progress of the medical field, plan the medical strategy for each new drug research and development pipeline for pharmaceutical companies, analyze the current status of unmet clinical needs of segmented populations, the current market competition status, explore opportunities for segmented populations, and provide clear product decisions.
[0087] It should be noted that the specific limitations of the disease treatment competition landscape information retrieval device can refer to the limitations of the disease treatment competition landscape information retrieval method above, which will not be repeated here. Each module in the above-mentioned disease treatment competition landscape information retrieval device can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the electronic device in the form of hardware, or can be stored in the memory of the electronic device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.
[0088] In some embodiments of the present application, the disease treatment competition landscape information retrieval device 300 can be implemented in the form of a computer program. The computer program can be used in Figure 4 The computer device shown in the figure is run. The memory of the computer device can store various program modules constituting the disease treatment competition pattern information retrieval device 300, for example, Figure 3 The entry acquisition module 310, information screening module 320 and information sending module 330 shown in the figure; the computer program composed of each program module enables the processor to execute the steps of the disease treatment competition landscape information retrieval method of each embodiment of the present application described in this specification. For example, Figure 4 The computer device shown can be Figure 3 The entry acquisition module 310 in the disease treatment competition landscape information retrieval device 300 shown executes step S201. The computer device can execute step S202 through the information screening module 320. The computer device can execute step S203 through the information sending module 330. Among them, the computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external computer device through a network connection. When the computer program is executed by the processor, a disease treatment competition landscape information retrieval method is implemented.
[0089] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0090] In some embodiments of the present application, a computer device is provided, comprising one or more processors; a memory; and one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the processor in the steps of the above-mentioned disease treatment competition landscape information retrieval method. The steps of the disease treatment competition landscape information retrieval method here may be the steps of the disease treatment competition landscape information retrieval method in each of the above-mentioned embodiments.
[0091] In some embodiments of the present application, a computer-readable storage medium is provided, which stores a computer program, and the computer program is loaded by a processor, so that the processor executes the steps of the above-mentioned disease treatment competition landscape information retrieval method. The steps of the disease treatment competition landscape information retrieval method here can be the steps of the disease treatment competition landscape information retrieval method in each of the above-mentioned embodiments.
[0092] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Any reference to memory, storage, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).
[0093] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0094] The above is a detailed introduction to a disease treatment competition landscape information retrieval method, device, computer equipment and computer-readable storage medium provided in an embodiment of the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of the present invention, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present invention.
Claims
1. A disease treatment competition landscape information retrieval method, characterized in that: include: Obtain user search terms for querying information on the competitive landscape of disease treatments; Filter out target information matching the user's search terms through a pre-built treatment pattern database; wherein the target information includes at least one of the following: medical evidence chain information, research and development progress information, combination information, predicted time to market, and special review information of the drug under segmented populations; The target information is sent to the client for display by the client.
2. The method according to claim 1, characterized in that Before obtaining the user search terms for querying the disease treatment competition landscape information, the method further includes: Obtain biomedical information, including clinical trial information, evidence-based results information, review and approval information, and guideline inclusion information; Performing key entity recognition and standardization processing on the biomedical information to obtain key entity information; Based on the key entity information, a population segmentation dictionary is constructed; wherein each population segmentation dictionary contained in the population segmentation dictionary is composed of at least two entity fields of indication, stage, pathology, biomarker, and baseline characteristics; Based on the dictionary of segmented populations, determine the medical evidence chain information, research and development progress information, combination information, predicted time to market and special review information of the drug under each segmented population; The medical evidence chain information, R&D progress information, regimen combination information, predicted time to market and special review information of drugs in various population segments are collected to construct the treatment pattern database.
3. The method according to claim 2, characterized in that The key entity identification and standardization processing of the biomedical information to obtain key entity information includes: Inputting the biomedical information into a trained entity recognition model and outputting initial entity information; Performing standardization processing on the initial entity information to obtain the key entity information; Among them, the key entity information includes but is not limited to at least one of the following: indications, staging, pathology, biomarkers, baseline characteristics, treatment thread, generic name of the drug, registration number, approval agency, and approval time; the trained entity recognition model is trained using preset label content, and the preset label content includes: indications, staging, pathology, biomarkers, baseline characteristics, treatment thread, and regimen elements.
4. The method according to claim 2, characterized in that The step of constructing a dictionary for segmented populations based on the key entity information includes: Counting the key entity information according to the entity fields constituting the segmented population to obtain statistical results; If the statistical result meets the preset conditions for adding a new segmented population, a pre-stored segmented population relationship table is retrieved and updated, wherein the segmented population relationship table is constructed based on the parent-child relationship of each segmented population; If the statistical result does not meet the population addition condition, the pre-stored segmented population relationship table is retrieved; With respect to the currently retrieved segmented population relationship table, a preset automatic mapping algorithm is used to automatically match each of the key entity information to the segmented population relationship table, so as to construct the segmented population dictionary.
5. The method according to any one of claims 2 to 4, characterized in that Based on the dictionary of segmented populations, the medical evidence chain information, research and development progress information, combination information, predicted time to market and special review information of the drug under each segmented population are determined, including: For each segmented population in the segmented population dictionary, extracting the associated treatment thread and drug generic name from the key entity information; The mutually related sub-populations, treatment threads and generic names of the drugs are used as data primary key information; For the data primary key information, according to the registration number in the clinical trial information, the clinical trial information, the evidence-based result information, the review and approval information and the guideline inclusion information are associated to obtain the medical evidence chain information; For the data primary key information, the highest R&D stage, as well as the milestone time and event status corresponding to the highest R&D stage are calculated to obtain the R&D progress information; Determine the scheme usage type for the data master key information as the scheme combination information; wherein the scheme usage type includes a single-drug scheme and a combination scheme; For the data primary key information, calculate the mean or median of the historical time to market of similar drugs as the predicted time to market; With respect to the data primary key information, at least one of the review document information, the regulatory agency release information and the news information is obtained, and when the preset special approval keywords are monitored, a special review tag is generated as the special review information.
6. The method according to claim 5, characterized in that The step of calculating the highest R&D stage, and the milestone time and event status corresponding to the highest R&D stage for the data primary key information to obtain the R&D progress information includes: For the data primary key information, based on the medical evidence chain information, determine the research and development stage corresponding to the clinical trial conducting area and / or the review area; wherein the research and development stages from high to low include at least: approval for marketing, application for marketing, phase III clinical trial, phase II / III clinical trial, phase II clinical trial, phase I / II clinical trial, phase I clinical trial; Based on the research and development stages corresponding to the clinical trial conducting regions and / or review regions, determine the highest research and development stage of the data primary key information in different regions; Determine the time of entering the highest research and development stage as the milestone time; The announcement status of the highest research and development stage is determined as the event status, and the research and development progress information is obtained by aggregation.
7. The method according to claim 5, characterized in that The step of determining a scheme attribute tag for the data primary key information as the scheme association information includes: determining biomedical information associated with the data primary key information; Inputting the biomedical information into a trained text classification model, and outputting a target sentence describing a treatment plan in the biomedical information; Inputting the target sentence into a trained pointer model, so that the pointer model extracts a treatment plan for the target sentence based on preset prompt information, and outputs a treatment plan corresponding to the prompt information; If the treatment regimen corresponding to the prompt information is a first treatment regimen that exists alone, determining the single-drug regimen as the combination information of the regimen; If the treatment plan corresponding to the prompt information is an associated second treatment plan, the combined plan is determined as the plan combination information.
8. A disease treatment competition landscape information retrieval device, characterized in that: include: A term acquisition module, used to acquire user search terms for querying information on the competitive landscape of disease treatment; An information screening module, for screening out target information matching the user's search terms through a pre-built treatment pattern database; wherein the target information includes at least one of the following: medical evidence chain information, research and development progress information, combination information, predicted time to market, and special review information of the drug under segmented populations; The information sending module is used to send the target information to the client for display by the client.
9. A computer device, characterized in that: The computer device comprises: one or more processors; A memory; and one or more applications, wherein the one or more applications are stored in the memory and configured to be executed by the processor to implement the disease treatment competition landscape information retrieval method described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and the computer program is loaded by a processor to execute the steps of the disease treatment competition landscape information retrieval method described in any one of claims 1 to 7.
Citation Information
Patent Citations
Medicine research and development competition data analysis method and device, electronic equipment and storage medium
CN113868488A
Disease information mining and retrieval method and device, electronic equipment and storage medium
CN114400099A
Biological medicine knowledge extraction method and device of decoder based on pointer network
CN116245178A
Combination therapy information analysis method and combination therapy information query method
CN116759107A
Flow disease information retrieval method and device
CN117573702A
