Umbilical cord mesenchymal stem cell donor matching database construction method based on HLA typing

Abstract

The invention relates to the technical field of biomedical information, in particular to an umbilical cord mesenchymal stem cell donor matching database construction method based on HLA (human leukocyte antigen) typing, which comprises the following steps of: recording and storing data of a donor in a database to finish initial construction of the database; regularly rechecking the donor data record stored in the database, and updating the state identifier of the donor data record when the rechecked data is changed by a preset degree; and receiving a matching request containing the human leukocyte antigen typing data of the target patient and the target disease type, performing donor screening in the database by using a preset comprehensive matching degree model based on the matching request, and optimizing parameters of the comprehensive matching degree model through a machine learning model based on the matching record and the clinical curative effect data corresponding to the matching record. According to the method, the accuracy and effectiveness of umbilical cord mesenchymal stem cell treatment are improved, and meanwhile, the system has continuous self-evolution ability.

Description

Technical Field

[0001] This invention relates to the field of biomedical information technology, and more specifically, to a method for constructing an umbilical cord mesenchymal stem cell donor matching database based on HLA typing. Background Technology

[0002] With the rapid development of regenerative medicine, cell therapy has become a revolutionary strategy for solving many intractable diseases. Among them, umbilical cord mesenchymal stem cells (UC-MSCs) are considered one of the most promising cell resources for clinical translation due to their outstanding advantages such as less ethical controversy, convenient acquisition, strong expansion capacity and multiple therapeutic mechanisms. They have been widely used in the fields of autoimmune diseases, tissue repair, and degenerative diseases.

[0003] However, the standardization and efficacy stability of UC-MSCs treatment remain key bottlenecks restricting its large-scale clinical promotion. Currently, when constructing UC-MSCs cell banks and selecting clinical donors, the industry generally follows the mature paradigm of hematopoietic stem cell transplantation, which relies heavily on human leukocyte antigen (HLA) typing matching in order to reduce immune rejection.

[0004] The existing technology has the following significant drawbacks: 1. Ignoring the core mechanism: The role of UC-MSCs mainly depends on their immune regulation and paracrine functions. They themselves have low immunogenicity, and excessive pursuit of HLA matching is often unnecessary, which may lead to the selection of "matched but ineffective" cells.

[0005] 2. Ignoring functional heterogeneity: UC-MSCs from different donors have huge functional differences, and existing databases lack assessment and recording of cellular functional efficacy, making it impossible to predict therapeutic effects.

[0006] Therefore, a method for constructing an umbilical cord mesenchymal stem cell donor matching database based on HLA typing is provided. Summary of the Invention

[0007] The purpose of this invention is to provide a method for constructing an HLA-based umbilical cord mesenchymal stem cell donor matching database to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention aims to provide a method for constructing an HLA-based umbilical cord mesenchymal stem cell donor matching database, comprising the following steps: S1. Collect source data from multiple umbilical cord mesenchymal stem cell donors, construct corresponding data records for each donor based on the source data, and store the donor data records in the database to complete the initial construction of the database; S2. Periodically review the donor data records stored in the database. When the reviewed data changes to a preset degree, update the status flag of the donor data record. S3. Receive a matching request containing the target patient's human leukocyte antigen typing data and the target disease type. Based on the matching request, use a preset comprehensive matching degree model to screen donors in the database. The comprehensive matching model generates a comprehensive matching score between donors and patients based on the first score of the first matching dimension and the second score of the second matching dimension, and generates a donor recommendation list based on the comprehensive matching score. S4. Collect and store matching records and their corresponding clinical efficacy data. Based on the matching records and their corresponding clinical efficacy data, optimize the parameters of the comprehensive matching degree model through machine learning models.

[0009] As a further improvement to this technical solution, in S1, the donor data record includes the donor's human leukocyte antigen typing information and a set of quantitative functional indicators.

[0010] As a further improvement to this technical solution, the set of quantitative functional indicators is numerical data obtained through a predefined standardized testing process, including at least the first type of indicators, the second type of indicators, and the third type of indicators. Among them, the first type of indicator can reflect the cell's immune regulation ability; the second type of indicator can reflect the cell's paracrine function; and the third type of indicator can reflect the cell's motility and migration ability.

[0011] As a further improvement to this technical solution, in step S2, when the verification data indicates that the key functional indicators of the donor have changed to a predetermined degree, the specific steps for updating the status identifier of the donor data record are as follows: S21. According to the predetermined time cycle, retrieve the stored donor data records from the database, and conduct sampling re-examination of the corresponding umbilical cord mesenchymal stem cells to obtain the current cell viability data and key functional indicator data. S22. Compare the current cell viability data and key functional indicator data with the previous verification data stored in the donor data record; S23. If the comparison results meet one of the following conditions, it is determined that the donor's key functional indicators have changed to a predetermined degree: Condition 1: The current cell viability drops below the first preset threshold; Condition 2: The numerical decay of any current key functional indicator exceeds its corresponding second preset threshold. S24. If it is determined that the key functional indicators of the donor have changed to a preset degree, the status identifier of the donor data record is updated to the functionally limited state, and the review data for this time is stored in the database.

[0012] As a further improvement to this technical solution, in S3, the first matching dimension is based on the degree of similarity of human leukocyte antigen typing information; the second matching dimension is based on the correlation between quantitative functional indicators and target disease types.

[0013] As a further improvement to this technical solution, in step S3, the screening of data records from multiple donors includes a preliminary screening step and a fine screening step. The initial screening step specifically includes: Using the target patient's human leukocyte antigen (HLA) typing data as the query condition, the database is used to filter out a set of candidate donors whose HLA typing information matches a third preset threshold. The fine sieving step is specifically as follows: For each donor in the candidate donor set, the comprehensive matching degree model is called to calculate the comprehensive matching degree score of the donor; The comprehensive matching degree model is as follows: ; in The overall matching score is defined as follows: S1 is the first score; S2 is the second score; W1 is the weight coefficient of the first matching dimension; and W2 is the weight coefficient of the second matching dimension.

[0014] As a further improvement to this technical solution, the second score is obtained based on a pre-constructed disease-indicator weight mapping table, wherein the disease-indicator weight mapping table uses disease type as the key and a set of functional indicators and their weight allocation schemes pre-set for the corresponding disease type as the value. The fine screening step obtains the weight allocation scheme corresponding to the target disease type by querying the disease-indicator weight mapping table, and calculates the second score by weighted summation of the quantitative functional indicators of the candidate donors.

[0015] As a further improvement to this technical solution, in step S3, a donor recommendation list is generated based on the comprehensive matching score. The generated donor recommendation list is a structured data object, which includes a donor identifier, a comprehensive matching score, and a score decomposition description field.

[0016] As a further improvement to this technical solution, in S4, the clinical efficacy data includes short-term efficacy indicators and long-term efficacy indicators. The short-term efficacy indicators include changes in the levels of disease-related specific biomarkers and improvements in clinical scoring scales detected at the first predetermined time point after treatment. The long-term efficacy endpoints include symptom recurrence rate, survival rate, and incidence of key adverse events recorded at a second predetermined time point after treatment. The clinical efficacy data are established in an immutable manner with the donor cell data records used, patient information, and treatment dosage and route.

[0017] As a further improvement to this technical solution, the specific steps in S4 for optimizing the comprehensive matching degree model based on the matching records and their corresponding clinical efficacy data using a machine learning model are as follows: S41. Obtain the feature data of the matching records, combine the feature data of each matching record into a feature vector, and form a training sample together with the corresponding clinical efficacy data. S42. Using a training dataset consisting of multiple training samples, train an efficacy prediction classification model, and after the model training is completed, obtain the importance contribution of each feature in the feature vector to the prediction result. S43. Based on the importance contribution, dynamically adjust the parameters in the comprehensive matching degree model. The parameters include at least the weight coefficient of the first matching dimension, the weight coefficient of the second matching dimension, and the weights in the disease-indicator weight mapping table.

[0018] Compared with the prior art, the beneficial effects of the present invention are as follows: This method for constructing an HLA-based umbilical cord mesenchymal stem cell donor matching database integrates cell functional efficacy indicators with disease-specific matching rules into an intelligent database system. It employs a comprehensive matching algorithm that prioritizes functional matching over HLA matching for donor selection and establishes a machine learning closed-loop optimization mechanism based on clinical efficacy feedback. This approach completely overcomes the limitations of traditional databases that rely solely on static matching through HLA typing, achieving a fundamental shift from "successful matching" to "promising therapeutic effect." It significantly improves the precision and effectiveness of umbilical cord mesenchymal stem cell therapy while enabling the system to continuously evolve. Attached Figure Description

[0019] Figure 1 This is a flowchart of the overall method of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention. Example

[0021] Please see Figure 1 As shown, this embodiment provides a method for constructing an HLA-based umbilical cord mesenchymal stem cell donor matching database, including the following steps: S1. Collect source data from multiple umbilical cord mesenchymal stem cell donors, construct corresponding data records for each donor based on the source data, and store the donor data records in the database to complete the initial construction of the database; In S1, the donor's data record includes the donor's human leukocyte antigen typing information and a set of quantitative functional indicators; A set of quantitative functional indicators is numerical data obtained through a predefined standardized testing process, including at least the first type of indicators, the second type of indicators, and the third type of indicators; Among them, the first type of indicator can reflect the cell's immune regulation ability; the second type of indicator can reflect the cell's paracrine function; and the third type of indicator can reflect the cell's motility and migration ability.

[0022] In this embodiment, the first type of indicators includes T cell proliferation inhibition rate and interferon-γ induced indoleamine 2,3-dioxygenase activity; T cell proliferation inhibition rate is used to measure the percentage of T cell proliferation that is inhibited by co-culturing UC-MSCs with activated T cells. This is the gold standard for measuring immunosuppressive capacity. Interferon-γ-induced indoleamine 2,3-dioxygenase activity was used to detect the activity of IDO secretion by UC-MSCs under inflammatory conditions. IDO is a key molecule mediating immune tolerance.

[0023] The second category of indicators includes the secretion levels of vascular endothelial growth factor, hepatocyte growth factor, and transforming growth factor-β. The secretion level of vascular endothelial growth factor (VEGF) is used to detect the concentration of VEGF in cell culture supernatant by methods such as ELISA, reflecting its ability to promote angiogenesis. The level of hepatocyte growth factor secretion is used to detect the concentration of HGF, reflecting its ability to resist fibrosis and promote tissue repair. Transforming growth factor-β secretion levels are used to detect TGF-β concentrations, which play a complex role in immune regulation and tissue repair.

[0024] The third category of indicators includes the wound healing rate in the scratch test and the number of cells that penetrate the membrane in the Transwell migration test. The scratch healing rate is used to measure the rate at which cells migrate to and “heal” a single-layer scratched area.

[0025] The Transwell migration assay is used to count the number of cells that have passed through the porous membrane, thus quantitatively assessing their active migration ability.

[0026] S2. Periodically review the donor data records stored in the database. When the reviewed data changes to a preset degree, update the status flag of the donor data record. In S2, when the review data indicates that the donor's key functional indicators have changed to a predetermined degree, the specific steps for updating the status flag of the donor data record are as follows: S21. According to the predetermined time cycle, retrieve the stored donor data records from the database, and conduct sampling re-inspection of the corresponding umbilical cord mesenchymal stem cells to obtain the current cell viability data and key functional indicator data; the predetermined time cycle is usually 3 to 6 months, which is synchronized with the cell bank's regular quality inspection cycle. Key functional indicators refer to the following: during the review, not all indicators may be tested repeatedly, but only "T cell proliferation inhibition rate" (first type of indicator) and "vascular endothelial growth factor (VEGF) secretion level" (second type of indicator) may be tested. Key functional indicators are representative indicators selected from "a set of quantitative functional indicators", and their decline is regarded as a barometer of the decline of overall cell function.

[0027] S22. Compare the current cell viability data and key functional indicator data with the previous verification data stored in the donor data record; if it is the first verification, compare it with the initial data. S23. If the comparison results meet one of the following conditions, it is determined that the donor's key functional indicators have changed to a predetermined degree: Condition 1: The current cell viability drops below a first preset threshold; the first preset threshold is set to 85% in this embodiment. If the retested viability is lower than this value, it indicates that the cell quality is no longer suitable for clinical use. Condition 2: The numerical decay of any current key functional indicator exceeds its corresponding second preset threshold; the second preset threshold is set differently for different functional indicators. In this embodiment, it is set to trigger an alarm if the value decreases by more than 30% compared to the initial value or the previous detection value. For example, if the initial T cell inhibition rate is 70% and it drops below 45% in a retest, it is determined to be functional decay.

[0028] S24. If it is determined that the key functional indicators of the donor have changed to a predetermined degree, the status identifier of the donor data record is updated to the functionally limited state, and the review data of this time is stored in the database. In this embodiment, updating the status identifier of the donor data record to a "functionally restricted" state specifically involves changing the status identifier from "available" to "functionally restricted." This status identifier will be read by the system in subsequent intelligent matching steps. Donors in the "functionally restricted" state may have their overall matching score multiplied by a penalty factor (such as 0.5), resulting in a significant drop in their ranking, or, under certain strict requirements, being directly excluded from the recommendation list.

[0029] S3. Receive a matching request containing the target patient's human leukocyte antigen typing data and the target disease type. Based on the matching request, use a preset comprehensive matching degree model to screen donors in the database. The comprehensive matching model generates a comprehensive matching score between donors and patients based on the first score of the first matching dimension and the second score of the second matching dimension, and generates a donor recommendation list based on the comprehensive matching score. In S3, the first matching dimension is based on the consistency of human leukocyte antigen typing information; the second matching dimension is based on the correlation between quantitative functional indicators and target disease types. The first matching dimension was obtained based on HLA typing nomenclature, directly comparing the alleles of HLA-A, B, and DRB1 loci in the donor and patient. Full match was S1=6 / 6=1.0; 5 / 6 match was S1=0.83; and 4 / 6 match was S1=0.67.

[0030] In S3, the screening of data records from multiple donors includes a primary screening step and a secondary screening step. The primary screening step specifically includes: Using the target patient's human leukocyte antigen (HLA) typing data as the query condition, a set of candidate donors with a matching degree of HLA typing information higher than a third preset threshold is selected from the database; in this embodiment, the third preset threshold is set to 4 / 6 matching. That is, the donor and the patient must match at least 4 alleles in the three loci of HLA-A, B, and DRB1 before entering the subsequent fine screening, which is a balance between the number of available donors and the difficulty of matching.

[0031] The specific steps of the fine sieve are as follows: For each donor in the candidate donor set, the comprehensive matching degree model is called to calculate the comprehensive matching degree score of the donor; The comprehensive matching degree model is as follows: ; in The overall matching score is defined as follows: S1 is the first score; S2 is the second score; W1 is the weight coefficient of the first matching dimension; W2 is the weight coefficient of the second matching dimension; and W2 > W1. The second score is obtained based on a pre-constructed disease-indicator weight mapping table. The disease-indicator weight mapping table uses disease type as the key and a set of pre-defined functional indicators and their weight allocation schemes corresponding to the disease type as the values. The fine screening step obtains the weight allocation scheme corresponding to the target disease type by querying the disease-indicator weight mapping table, and performs a weighted summation of the quantitative functional indicators of the candidate donors to calculate the second score.

[0032] Traditional HLA matching databases assume all eligible cells are "equivalent," meaning that selecting a cell that excels at immunosuppression may not be as effective as selecting a cell that excels at secreting cartilage repair factors and has strong migration capabilities for an arthritis patient.

[0033] This approach uses a "disease-indicator weight mapping table" to encode medical knowledge (which diseases require which cell functions) into the matching algorithm. For GVHD (graft-versus-host disease) patients, it prioritizes recommending cells with high scores on the "immunosuppression" indicator in the S2 score; for myocardial infarction patients, it prioritizes recommending cells with strong secretion capabilities of pro-angiogenic factors such as VEGF and HGF. This ensures that the recommended cells are highly functionally aligned with the patient's pathological needs, significantly increasing the probability of treatment success.

[0034] In S3, a donor recommendation list is generated based on the overall matching score. The generated donor recommendation list is a structured data object, which includes a donor identifier, an overall matching score, and a score decomposition description field. The score decomposition description field is used to explain to the client the advantages of each recommended donor in the first matching dimension and the second matching dimension.

[0035] The donor recommendation list includes a score breakdown description field, increasing the transparency and interpretability of the system's decision-making process and providing clinical decision support for doctors, rather than simply providing a cold ranking list.

[0036] This allows doctors to clearly understand whether a donor is recommended because of its high HLA match or because it performs well in a specific function, thus enabling them to make a more informed final choice. For example, the top-ranked donor may be an HLA 5 / 6 match with extremely strong function (high S2 score), while the second-ranked donor may be an HLA 6 / 6 match with moderate function (high S1 score). Doctors can make personalized assessments based on the patient's specific condition, immune status, and other factors, with the assistance of the system.

[0037] S4. Collect and store matching records and their corresponding clinical efficacy data. Based on the matching records and their corresponding clinical efficacy data, optimize the parameters of the comprehensive matching degree model through machine learning models.

[0038] In S4, clinical efficacy data include short-term efficacy indicators and long-term efficacy indicators; Short-term efficacy endpoints include changes in disease-related specific biomarker levels and improvements in clinical scoring scales, measured at the first predetermined time point after treatment. The first predetermined time point refers to the window period used to evaluate short-term efficacy. The setting of this time point needs to comprehensively consider the natural course of the disease, the expected onset time of treatment, and routine clinical follow-up points. In this embodiment, the first predetermined time point is from week 4 to week 12 after treatment.

[0039] Long-term efficacy endpoints include symptom recurrence rate, survival rate, and incidence of key adverse events recorded at the second predetermined time point after treatment; The second predetermined time point refers to the endpoint observation period used to assess long-term efficacy and safety. This time point aims to confirm the durability and long-term effects of the treatment. In this embodiment, the second predetermined time point is from 6 months to 24 months after treatment.

[0040] Clinical efficacy data are linked in an immutable manner to the donor cell data records used, patient information, and treatment dosage and route.

[0041] The purpose of establishing an immutable link is to ensure the integrity and traceability of the data chain, providing high-quality and reliable training data for machine learning. In practice, blockchain technology is used to upload the hash values ​​of "donor cell identifiers (such as serial numbers)," "patient anonymity identifiers," "treatment dosage and route," and "efficacy data" to the chain, forming an immutable data package.

[0042] In S4, the specific steps for optimizing the comprehensive matching degree model using a machine learning model based on the matched records and their corresponding clinical efficacy data are as follows: S41. Obtain the feature data of the matching records, combine the feature data of each matching record into a feature vector, and form a training sample together with the corresponding clinical efficacy data; the feature data includes the first score and second score of the corresponding donor and patient, as well as the values ​​of multiple underlying quantitative functional indicators that constitute the second score; the clinical efficacy data is converted into a binary efficacy result label; each historical matching record is associated with a clinical efficacy data. A binary classification efficacy outcome label of 1 indicates that the treatment is effective; a binary classification efficacy outcome label of 0 indicates that the treatment is ineffective.

[0043] S42. Using a training dataset consisting of multiple training samples, train an efficacy prediction classification model, and after the model training is completed, obtain the importance contribution of each feature in the feature vector to the prediction result. The dataset is (X, Y), where X is the feature vector and Y is the binary classification label. The efficacy prediction classification model is a binary classification model, and the purpose of the model is to learn... , which is the probability that the treatment is "effective" given a matching feature X.

[0044] S43. Based on the importance contribution, dynamically adjust the parameters in the comprehensive matching degree model. The parameters include at least the weight coefficient of the first matching dimension, the weight coefficient of the second matching dimension, and the weights in the disease-indicator weight mapping table. The optimized weight coefficients of the first matching dimension, the weight coefficients of the second matching dimension, and the new disease-indicator weight mapping table are applied to the next matching recommendation.

[0045] New matches generate new clinical efficacy data, which are then added to the historical database, triggering a new cycle of "data preparation → model training → importance analysis → strategy optimization". Through this continuous closed loop, the system can adapt to advancements in medical understanding and changes in cell characteristics, becoming increasingly accurate.

[0046] A specific example: finding a suitable UC-MSCs donor for a patient with severe graft-versus-host disease (GVHD); First, 1,000 UC-MSCs were collected, and each sample was subjected to high-resolution HLA typing. Various functional indicators, such as T cell inhibition rate, VEGF, and HGF secretion levels, were also detected and stored in a database. Quarterly, random samples of stored cells are re-examined. It was found that the T cell inhibition rate of cell #123 had decreased from the initial 75% to 40%, so its status was updated to "functionally restricted". Furthermore, doctors submit matching requests for GVHD patients through the system, which filters out 200 candidate donors who are HLA ≥ 4 / 6 compatible with the patient. The system calls the "Disease-Indicator Weight Mapping Table" and sets the weights for "GVHD" as follows: T cell inhibition rate weight is 0.7, IDO activity weight is 0.3 (other indicators weight is 0).

[0047] The system calculates the overall matching score for these 200 donors. Because W2 > W1 in the algorithm, a donor #456 with an HLA 5 / 6 match but a high inhibition rate of 80% will have a much higher score than a donor #789 with an HLA 6 / 6 match but a inhibition rate of only 50%. Cell #123, whose status is "functionally limited," was automatically downweighted by the system, even though the initial data was good.

[0048] Ultimately, the system returned a recommendation list, with donor #456 ranked first. The "Score Breakdown Explanation" field showed: "Reason for recommendation: HLA 5 / 6 match (good), core advantage: extremely strong immunosuppressive function (T cell suppression rate: 80%)."

[0049] Based on this information, the doctor approved and selected #456 cells for treatment.

[0050] After treatment, the doctor will transmit the patient's efficacy data (such as the complete relief of the patient's rash and diarrhea 28 days after treatment, which is considered "effective") back through the system.

[0051] The system records successful cases of "effective treatment of GVHD using cells with high inhibition rates." With sufficient data accumulation, the machine learning model may discover that for GVHD, the weight of "T cell inhibition rate" (0.7) or even higher can be increased. The system will then automatically optimize its matching strategy to recommend cells with greater efficacy guarantees for GVHD patients in the future.

[0052] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A method for constructing a database of umbilical cord mesenchymal stem cell donor matching based on HLA typing, characterized in that, The method comprises the following steps: S1, collecting source data of a plurality of umbilical cord mesenchymal stem cell donors, constructing a corresponding data record for each donor according to the source data, and storing the data record of the donor in the database to complete the initialization construction of the database; S2, periodically reviewing the donor data records stored in the database, and updating the state identifier of the donor data record when the reviewed data changes by a preset degree; S3, receiving a matching request containing human leukocyte antigen typing data of a target patient and a target disease type, and screening donors in the database based on the matching request using a preset comprehensive matching degree model; Wherein, the comprehensive matching degree model generates a comprehensive matching degree score between the donor and the patient according to the first score of the first matching dimension and the second score of the second matching dimension, and generates a donor recommendation list according to the comprehensive matching degree score; S4, collecting and storing matching records and their corresponding clinical efficacy data, and optimizing the parameters of the comprehensive matching degree model through a machine learning model based on the matching records and their corresponding clinical efficacy data.

2. The method for constructing an HLA-based umbilical cord mesenchymal stem cell donor matching database according to claim 1, characterized in that: In S1, the data record of the donor includes human leukocyte antigen typing information of the donor and a set of quantitative functional indicators.

3. The method of claim 2, wherein the HLA-typing based umbilical cord mesenchymal stem cell donor matching database construction method is characterized by: The set of quantitative functional indicators are numerical data obtained through a pre-defined standardized detection process, and at least include a first type of indicator, a second type of indicator and a third type of indicator; Wherein, the first type of indicator can reflect the immune regulation ability of the cell; the second type of indicator can reflect the paracrine function of the cell; and the third type of indicator can reflect the movement and migration ability of the cell.

4. The method for constructing an HLA-based umbilical cord mesenchymal stem cell donor matching database according to claim 3, characterized in that: In S2, when the reviewed data indicates that the key functional indicators of the donor have changed by a preset degree, the state identifier of the donor data record is updated as follows: S21, according to a predetermined time period, the stored donor data record is retrieved from the database, and the corresponding inventory umbilical cord mesenchymal stem cells of the donor are sampled and rechecked to obtain the current cell viability data and key functional indicator data; S22, compare the current cell viability data and key functional indicator data with the last review data stored in the donor data record; S23, if the comparison result meets one of the following conditions, it is determined that the key functional indicators of the donor have changed by a preset degree: Condition one: the current cell viability decreases below a first preset threshold; Condition two: the numerical attenuation amplitude of any current key functional indicator exceeds the corresponding second preset threshold; S24, if it is determined that the key functional indicators of the donor have changed by a preset degree, the state identifier of the donor data record is updated to a functionally limited state, and the review data of this time is stored in the database.

5. The method for constructing a database of HLA-typed umbilical cord mesenchymal stem cell donors for matching, according to claim 4, characterized in that: In S3, the first matching dimension is based on the compatibility of human leukocyte antigen typing information; and the second matching dimension is based on the correlation degree of quantitative functional indicators and target disease types.

6. The method for constructing an HLA-based umbilical cord mesenchymal stem cell donor matching database according to claim 5, characterized in that: In S3, the screening in the data records of a plurality of donors includes a preliminary screening sub-step and a refined screening sub-step; Wherein, the preliminary screening sub-step is specifically: Filtering, from the database, a candidate donor set with human leukocyte antigen typing data of a target patient as a query condition, all human leukocyte antigen typing information with a matching degree higher than a third preset threshold; The fine screening step is specifically: For each donor in the candidate donor set, calling the comprehensive matching degree model to calculate the comprehensive matching degree score of the donor; The comprehensive matching degree model is specifically: ； wherein is a comprehensive matching score, S1 is a first score; S2 is a second score; W1 is a weight coefficient of a first matching dimension; and W2 is a weight coefficient of a second matching dimension.

7. The method of claim 6, wherein the HLA-typing based umbilical cord mesenchymal stem cell donor matching database construction method is characterized by: The second score is obtained based on a pre-constructed disease-index weight mapping table, the disease-index weight mapping table taking a disease type as a key and a preset set of functional indexes and their weight distribution scheme corresponding to the disease type as a value; The fine screening step obtains the weight distribution scheme corresponding to the target disease type by querying the disease-index weight mapping table, and performs weighted summation on the quantitative functional indexes of the candidate donor to calculate the second score.

8. The method of claim 7, wherein the HLA-typing based umbilical cord mesenchymal stem cell donor matching database construction method is characterized by: In S3, a donor recommendation list is generated according to the comprehensive matching degree score, wherein the generated donor recommendation list is a structured data object, which includes a donor identifier, a comprehensive matching degree score, and a score decomposition description field.

9. The method of claim 8, wherein the HLA-typing based umbilical cord mesenchymal stem cell donor matching database construction method is characterized by: In S4, the clinical efficacy data includes short-term efficacy indicators and long-term efficacy indicators; The short-term efficacy indicators include changes in the levels of specific biomarkers related to the disease and improvements in clinical score scales detected within a first predetermined time point after treatment; The long-term efficacy indicators include the recurrence rate, survival rate, and incidence of key adverse events recorded within a second predetermined time point after treatment; The clinical efficacy data is associated with the used donor cell data record, patient information, and treatment dose and route in a non-tamperable manner.

10. The method of claim 9, wherein the HLA-typing based umbilical cord mesenchymal stem cell donor matching database construction method is characterized by: In S4, based on the matching record and its corresponding clinical efficacy data, the specific steps of optimizing the comprehensive matching degree model by a machine learning model are as follows: S41, obtaining feature data of the matching record, combining the feature data of each matching record into a feature vector, and together with the corresponding clinical efficacy data to form a training sample; S42, using a training data set composed of a plurality of training samples to train an efficacy prediction classification model, and after the model training is completed, obtaining the importance contribution of each feature in the feature vector to the prediction result; S43, based on the importance contribution, dynamically adjusting the parameters in the comprehensive matching degree model, the parameters at least including the weight coefficient of the first matching dimension, the weight coefficient of the second matching dimension, and the weight in the disease-index weight mapping table.

Images