A method for monitoring immune status and evaluating prognosis in patients with thymoma after surgery

Abstract

The present invention belongs to the technical field of postoperative immune status assessment for thymoma patients, and specifically relates to a method for postoperative immune status monitoring and prognostic assessment for thymoma patients. The invention achieves dynamic tracking of the patient's immune status by accurately monitoring and assessing the immune status of thymoma patients at different postoperative periods. By collecting blood samples from patients at different postoperative periods and detecting the proportion of T cell subpopulations therein, the recovery of the patient's postoperative immune function can be reflected. At the same time, the present invention also effectively eliminates invalid sample parameters by screening, classifying and evaluating sample parameters, thereby improving the accuracy of the patient's immune status assessment, thereby providing doctors with a more reliable assessment basis and helping doctors formulate more scientific and reasonable treatment plans. In addition, the prognostic risk level can be determined based on the patient's immune deviation amount, providing doctors with a more reliable prognostic assessment basis.

Description

Technical Field

[0001] The present invention belongs to the technical field of postoperative immune status assessment for thymoma patients, and particularly relates to a method for postoperative immune status monitoring and prognosis assessment for thymoma patients. Background Art

[0002] With the popularization of thymoma surgery and the increase in patients' rehabilitation needs, accurate monitoring of the postoperative immune status and prognostic evaluation of thymoma patients have become particularly important. The three-port procedure is a new method of thymoma resection. It removes thymoma through three small incisions, which has less trauma and shorter recovery time than traditional thoracotomy. After thymoma patients undergo this surgical treatment, monitoring of their immune status and prognostic evaluation are crucial for their recovery and subsequent treatment. The traditional method is generally through regular review, and then relies on the doctor's clinical experience and the patient's clinical manifestations for evaluation. However, this method has the problems of strong subjectivity and insufficient accuracy. Therefore, it is particularly important to develop a more objective and accurate method for evaluating the postoperative immune status of thymoma patients.

[0003] Most of the evaluation methods in the existing technology rely on the patient's real-time sampling indicators, and then combine them with machine algorithms to perform real-time evaluation. However, due to individual differences among patients, the actual evaluation process is difficult to apply to complex differences. In particular, the quantification of individual differences is currently difficult to implement, resulting in the evaluation results being too dependent on reference values. In addition, due to the existence of individual differences, the change pattern of postoperative immune status is actually difficult to predict. In addition, the immune status of postoperative patients is constantly changing, and the reference value of a large amount of historical data is limited. Conventional interval evaluation methods are difficult to accurately reflect the patient's current immune status and may even produce misleading evaluation results. Therefore, the present invention proposes a method for monitoring the immune status and prognosis evaluation of thymoma patients after surgery, aiming to solve the above problems. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for monitoring the immune status and evaluating the prognosis of thymoma patients after surgery, which can realize targeted monitoring and evaluation of the patient's postoperative immune status and improve the accuracy and reliability of prognosis evaluation through dynamic monitoring.

[0005] The technical solutions adopted by the present invention are as follows:

[0006] Based on preoperative immune status data, including cell subset ratios, patients are clustered using a cluster analysis algorithm to construct a multi-level immune status feature cluster. Based on the feature cluster, each patient's immune status category and individualized fluctuation threshold are determined. Clustering includes preliminary clustering using a coarse-grained clustering algorithm and refined clustering using a density clustering algorithm.

[0007] Collect immune data samples from patients at consecutive time points after surgery and use them to assess the patient's immune status. The assessment process determines whether the immune status is normal or abnormal based on an individualized fluctuation threshold.

[0008] Under normal conditions and without postoperative complications, the corresponding immune data samples are collected and summarized as a normal state parameter set. If there are postoperative complications, it is re-marked as an abnormal state;

[0009] Under abnormal conditions, the early warning mechanism is triggered, and the amount of immune deviation is collected simultaneously. Combined with postoperative complications, the prognostic risk level is determined.

[0010] Furthermore, based on the preoperative immune status data, the patients are clustered by a cluster analysis algorithm to construct a multi-level immune status characteristic cluster, including:

[0011] The proportion of T cell subsets was detected from the patient's preoperative blood samples and recorded as preoperative immune status data;

[0012] Transfer learning is performed on preoperative immune status data through pre-trained immune feature encoders to generate high-order feature vectors. Dimensionality reduction tools are used to map high-order feature vectors to low-dimensional visualization space to obtain mapping results.

[0013] A coarse-grained clustering algorithm was used to preliminarily group the mapping results and determine the distribution of the preset immune activity categories;

[0014] Apply density clustering algorithm to refine the mapping results and set dynamic threshold parameters to identify small subgroups not covered by coarse-grained clustering algorithm;

[0015] Based on the preliminary clustering results of the coarse-grained clustering algorithm and the tiny subgroups identified by the density clustering algorithm, a multi-level immune status characteristic distribution group was formed.

[0016] Furthermore, after collecting the immune data samples of the patient at consecutive time points after surgery, the immune data samples need to be preprocessed. The preprocessing method includes:

[0017] The immune data samples are timestamped, and similar data at adjacent time points are merged through fluctuation parameter analysis to generate standardized time series parameters. Valid samples are screened from the standardized time series parameters, and interference numbers are eliminated in combination with antibiotic usage markers.

[0018] Furthermore, the step of obtaining valid samples includes:

[0019] Samples are screened based on standardized timing parameters, and the screening results are recorded as verification parameters;

[0020] According to the standardized time series parameters, the immune data samples are grouped in pairs according to the arrangement order to obtain multiple groups of sample evaluation data groups;

[0021] Based on the verification parameters, the immune data samples in each sample evaluation data group are verified and the verification pass rate of the sample evaluation data group is determined;

[0022] Determine the length of the selection period for the immune data sample based on the pass rate of each verification;

[0023] The last node of the timestamp in the immune data sample is used as the reference node, and the reference node is reversely offset according to the length of the selected period to obtain the sample period;

[0024] Immunity data samples covered within the sample period are recorded as valid samples, and samples outside the sample period are recorded as invalid samples.

[0025] Furthermore, the step of determining the verification pass rate of the sample evaluation data set includes:

[0026] Get the verification condition parameters;

[0027] Obtain a calibration algorithm, and input the calibration condition parameters and the immune data sample in the sample evaluation data group into the calibration algorithm to obtain a calibration bias;

[0028] Obtaining a verification threshold and comparing the verification threshold with a verification bias; and when the verification bias is greater than the verification threshold, indicating that the corresponding sample evaluation data group passes the verification; otherwise, indicating that the corresponding sample evaluation data group fails the verification;

[0029] The percentage of sample evaluation data groups that passed the verification in all sample evaluation data groups is counted and recorded as the verification pass rate of the sample evaluation data group.

[0030] Furthermore, the determination of the length of the selection period of the immune data sample based on the verification pass rate of each sample evaluation data group includes:

[0031] Obtain the verification pass rate and record it as the parameter to be evaluated;

[0032] Obtaining an evaluation interval, wherein multiple evaluation intervals are set, and each evaluation interval corresponds to an initial selection period;

[0033] Compare the parameter to be evaluated with the evaluation interval, match the corresponding initial selection period, and collect the number of immune data samples of the sample evaluation data group within the initial selection period, and record it as the selection condition parameter;

[0034] Obtaining an evaluation threshold and comparing the evaluation threshold with a selection condition parameter;

[0035] When the selection condition parameter is greater than the evaluation threshold, the initial selection period is directly determined as the selection period of the sample parameter. Otherwise, the length of the initial selection period is extended, and the selection condition parameters corresponding to the sample evaluation data group within the extended initial selection period are collected again until the selection condition parameter is greater than the evaluation threshold, and the extended initial selection period at this time is determined as the length of the selection period of the sample parameter.

[0036] Furthermore, the evaluation of the patient's immune status through the immune data sample includes:

[0037] After determining the valid samples, the valid samples are input into the evaluation function, and the output result of the evaluation function is recorded as the evaluation condition parameter;

[0038] Obtaining the state classification interval, adjusting the state classification interval with reference to the individualized fluctuation threshold, and then comparing it with the evaluation condition parameter;

[0039] When the evaluation condition parameter is in the adjusted status classification interval, it indicates that the patient's immune status is normal, and the patient's immune status is recorded as normal;

[0040] When the evaluation condition parameter exceeds the adjusted status classification interval, it indicates that the patient's immune status is abnormal, and the patient's immune status is recorded as abnormal.

[0041] Furthermore, in the abnormal state, the early warning mechanism is triggered, and the immune deviation amount is collected simultaneously, and the prognostic risk level is determined in combination with the postoperative complications, including:

[0042] Obtaining the immune deviation value, where the immune deviation value is the difference between the evaluation condition parameter and the upper and lower limits of the adjusted status classification interval, and correcting the immune deviation value by combining the postoperative complications through weighted fusion;

[0043] Obtaining a grading interval, where multiple grading intervals are set, and each grading interval corresponds to a prognostic risk level;

[0044] Match the corrected immune deviation amount with the grading interval to obtain the corresponding prognostic risk level;

[0045] Among them, the higher the prognostic risk level, the further the patient's immune status deviates from the normal range.

[0046] The present invention also provides a system for monitoring the immune status and evaluating the prognosis of patients after surgery for thymoma. The system uses the method for monitoring the immune status and evaluating the prognosis of patients after surgery for thymoma, comprising:

[0047] The preoperative clustering module is used to cluster patients using a cluster analysis algorithm, construct a multi-level immune status feature cluster, and determine each patient's immune status category and individualized fluctuation threshold based on the feature cluster. Clustering includes preliminary clustering using a coarse-grained clustering algorithm and refined clustering using a density clustering algorithm.

[0048] The sample collection module is used to collect immune data samples of patients at consecutive time points after surgery;

[0049] The immune assessment module evaluates the patient's immune status using immune data samples. The assessment process determines whether the immune status is normal or abnormal based on individualized fluctuation thresholds.

[0050] The aggregation module collects corresponding immune data samples when the patient is in a normal state and there are no postoperative complications, and summarizes them into a normal state parameter set. If there are postoperative complications, the data is re-marked as an abnormal state.

[0051] The alarm module is used to trigger the early warning mechanism under abnormal conditions, and simultaneously collect the amount of immune deviation, and determine the prognostic risk level in combination with postoperative complications.

[0052] The present invention further provides an electronic device, comprising:

[0053] at least one processor;

[0054] and a memory communicatively coupled to the at least one processor;

[0055] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for monitoring the postoperative immune status and evaluating the prognosis of thymoma patients.

[0056] The technical effects achieved by the present invention are:

[0057] The present invention groups the preoperative immune status of thymoma patients so that the monitoring and evaluation process is combined with the individual differences of patients, thereby making the evaluation results more accurate. In addition, through the dynamic tracking of the immune status, the recovery of the patient's immune function after surgery can be reflected. At the same time, the present invention also effectively eliminates invalid immune data samples by screening, classifying and evaluating immune data samples, and improves the accuracy of the patient's immune status assessment. It can also dynamically adjust the time period length so that the data evaluation is more in line with the patient's own condition, thereby providing doctors with a more reliable evaluation basis and helping doctors to formulate more scientific and reasonable treatment plans. In addition, the present invention can also determine the prognostic risk level according to the patient's immune deviation amount, providing doctors with a more reliable prognostic evaluation basis, thereby improving the scientificity and practicality of postoperative immune status monitoring and prognostic evaluation of thymoma patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic flow chart of the method of the present invention;

[0059] Figure 2 It is a schematic diagram of the system modules of the present invention;

[0060] Figure 3 It is a schematic structural diagram of an electronic device of the present invention. DETAILED DESCRIPTION

[0061] See also Figure 1 As shown, the present invention provides a method for monitoring the immune status and evaluating the prognosis of patients with thymoma after surgery, comprising:

[0062] S1. Based on preoperative immune status data, including cell subset ratios, patients are clustered using a cluster analysis algorithm to construct a multi-level immune status feature cluster. Based on the feature cluster, the immune status category and individualized fluctuation threshold for each patient are determined. Clustering includes preliminary clustering using a coarse-grained clustering algorithm and refined clustering using a density clustering algorithm.

[0063] In step S1, when evaluating the postoperative immune status of thymoma patients, the cell subset ratio (CD4+ / CD8+ ratio, Treg cell ratio) is first detected from the patient's preoperative samples, including but not limited to blood samples. The data is then summarized as preoperative immune status data. It should be noted that the cell subset ratio is the influencing factor selected in this application. In order to better understand this scheme, in the actual evaluation process, the immune status data can also introduce thymoma-specific cytokines (IL-10, TGF-β) and inflammatory markers (CRP, IL-6) concentrations, etc., in order to better group the patients according to their preoperative status. Of course, other factors that can reflect the patient's immune status can also be used, which will not be described in detail here.

[0064] After acquiring immune status data, transfer learning can be performed on preoperative immune status data using a pretrained immune feature encoder to generate high-order feature vectors. This can be achieved using dimensionality reduction tools, such as Unified Mapping (UMAP). UMAP reduces 128-dimensional features to 2 dimensions by preserving the local structure of the data, generating a scatter plot. This visual scatter plot helps to initially assess differences in immune status distribution and intuitively reveals underlying population characteristics. The high-order feature vectors are then mapped to a low-dimensional visualization space to obtain the mapping result. It should be noted that using a pretrained immune feature encoder for transfer learning is a common approach. The basic principle is that the pretrained encoder, by learning on large-scale immune-related data, can extract deep features related to immune activity. Deep neural networks (DNNs) are preferred as pretrained models. For example, a dataset containing preoperative blood samples from 1,000 patients, each containing data on multiple immune cell markers, can be used. Through transfer learning, this raw data is fed into the encoder to generate high-order feature vectors, such as a 128-dimensional feature vector for each patient sample. These vectors capture the complex patterns of the immune system and lay the foundation for subsequent analysis. In addition, other methods can be used, such as the Spearman correlation coefficient to screen the core feature vectors, map the core feature vectors to a two-dimensional space using the t-SNE algorithm, and then use the Euclidean distance method to determine the distribution of the immune features. Of course, the above methods are all relatively mature technologies in current machine learning algorithms. They are only used here as a data processing method, so their basic principles are not supplemented in too much detail. Of course, other algorithms or models that can achieve the above data processing purposes can also be applied in this application.

[0065] Then, a coarse-grained clustering algorithm is used to preliminarily group the mapping results to determine the preset immune activity category distribution;

[0066] A density clustering algorithm is used to refine the mapping results and set dynamic threshold parameters to identify small subgroups not covered by the coarse-grained clustering algorithm;

[0067] The preliminary clustering results of the coarse-grained clustering algorithm and the tiny subgroups identified by the density clustering algorithm are integrated to form a multi-level immune status characteristic distribution group.

[0068] For clustering, the coarse-grained clustering algorithm in this application adopts the K-means algorithm, and the patients are preliminarily clustered by presetting multiple categories, wherein the clustering basis of multiple categories can be classifications such as immunosuppression type, and the immunosuppression type clustering basis can be CD4+ / CD8+ ratio less than the preset value, and the Treg cell ratio is greater than the preset ratio. Of course, in addition to this method, other types such as high inflammation type, local immunosuppression type or other types of clustering can also be clustered, wherein the clustering method can be high inflammation type for IL-6, CRP significantly higher than the normal range, local immunosuppression type, IL-10, TGF-β are all higher than the threshold. The specific clustering type and range or threshold setting can be selected and set based on clinical or historical immune data, and will not be described in detail here. The purpose of using the K-means algorithm is to simplify the clustering method and to quickly divide the main groups, but it may ignore small-scale subgroups. Therefore, this application also uses the HDBSCAN density clustering algorithm to further identify small subgroups. By setting a minimum sample number threshold (the minimum sample number threshold is set to the total number of samples / fixed value, which is not specifically limited here), based on density differences, small clusters not identified by the K-means algorithm are identified. For example, a unique subgroup of 20 samples is identified in a certain category group, which may correspond to a specific immunosuppressive state. The advantage of this method is that it can capture subtle differences and improve clustering accuracy. By integrating the K-means clustering results with the subgroups identified by HDBSCAN, a complete immune status feature distribution group is formed. It not only retains the integrity of the main classification, but also reveals hidden minor features, which is for subsequent individual differences and improves the evaluation effect of prognostic immune ability.

[0069] In the above method, based on the clustering results, the patient's immune status category is divided into, such as immunosuppressive type, or multiple types coexist. When multiple types coexist, each status category can be scored, and then weighted fusion is performed based on each immune status category (when there is only a single cluster, the score is directly output, and when multiple clusters are weighted), and finally, based on the preset score range, the standard threshold ratio of the patient's subsequent score is determined. The standard threshold ratio is a reference to the standard threshold. For example, patients with high inflammation are more likely to develop inflammation, so their postoperative immune indicators are difficult to reach normal standards, but they are still evaluated normally, which easily leads to them being in an alarm state all the time. However, because they are a high inflammation type cluster, their postoperative verification risk will be higher than that of ordinary patients. Therefore, under the premise of appropriately lowering their alarm threshold (such as 0.9 times the standard threshold), more attention is paid to the trend changes of their postoperative immune samples, that is, while lowering the threshold, the fluctuation monitoring of the trend change is increased (the fluctuation change rate is set to 0.9 times the standard change rate), thereby realizing an immune monitoring method tailored to individual differences.

[0070] S2. Collect immune data samples at consecutive time points after surgery (it is best if the data type is consistent with the preoperative immune status data. Other data can also be introduced, such as IL-6 concentration in addition to CD4+ and CD8+ cell counts, which are not described in detail here). Evaluate the patient's immune status through the immune data samples. The evaluation process determines whether the immune status is normal or abnormal based on an individualized fluctuation threshold.

[0071] In the above step S2, after collecting the immune data samples of the patient at consecutive time nodes after surgery, the immune data samples need to be preprocessed. The preprocessing method includes:

[0072] The immune data samples are timestamped, and similar data from adjacent time points are merged through fluctuation parameter analysis to generate standardized time series parameters. Valid samples are screened from the standardized time series parameters, and interfering data are eliminated in combination with antibiotic usage marks. The process of adding timestamps is a relatively conventional data processing method. During postoperative immune monitoring, if patients use antibiotics due to infection or other reasons, it may cause short-term abnormal fluctuations in immune indicators (such as CD4+ / CD8+ ratio and Treg ratio). This step excludes external drug interference by marking the data during the antibiotic use time period, avoiding misjudgment of the patient's true immune status. Of course, the above-mentioned antibiotics do not constitute a limitation on other influencing factors. During the monitoring period, other factors that may affect the fluctuation of immune indicators can also be excluded by marking the use time period to reduce external drug interference.

[0073] The specific steps to obtain valid samples include:

[0074] Samples are screened based on standardized timing parameters, and the screening results are recorded as verification parameters. Standard timing parameters are used to arrange immune data samples in chronological order.

[0075] According to the standardized time series parameters, the immune data samples are grouped in pairs according to the arrangement order to obtain multiple groups of sample evaluation data groups;

[0076] Based on the verification parameters, the immune data samples in each sample evaluation data group are verified and the verification pass rate of the sample evaluation data group is determined;

[0077] Determine the length of the selection period for the immune data sample based on the pass rate of each verification;

[0078] The last node of the timestamp in the immune data sample is used as the reference node, and the reference node is reversely offset according to the length of the selected period to obtain the sample period;

[0079] Immunity data samples covered within the sample period are recorded as valid samples, and samples outside the sample period are recorded as invalid samples.

[0080] Specifically, in the above steps, when screening valid samples, it is necessary to screen according to standardized timing parameters. By arranging the immune data samples in order, the arranged immune data samples can be screened using machine algorithms or processing functions. In a preferred embodiment of the present invention, a predefined screening function can be used. The screening process requires inputting the sequentially arranged immune data samples into the screening function one by one, and the basis for subsequent verification, i.e., the verification condition parameters, can be obtained. The expression of the screening function can be: , where Indicates the verification condition parameters, represents the number of immune data samples, represents the length of the period covering the immunization data sample, and Represent the immune data samples under adjacent nodes, and then group them in pairs according to the arrangement order of the immune data samples to form multiple groups of sample evaluation data groups. Each group of data contains two adjacent immune data samples for corresponding verification processing. According to the obtained verification condition parameters, the immune data samples in each sample evaluation data group are verified to determine the verification pass rate of the sample evaluation data group, providing a quantitative basis for the subsequent immune data sample selection. On this basis, according to the verification pass rate of each sample evaluation data group, the length of the selection period of the immune data sample is scientifically and reasonably determined to ensure that the selection period can cover valid data and exclude invalid data. Subsequently, the last node of the timestamp in the immune data sample is used as the reference node, and the reference node is reversely offset according to the determined selection period length, thereby delineating the range of the sample period. Finally, all immune data samples covered in the sample period are recorded in detail as valid immune data samples. At the same time, the immune data samples outside the sample period are also recorded in detail as invalid immune data samples, completing the distinction between the validity and invalidity of the immune data samples.

[0081] Furthermore, the step of determining the verification pass rate of the sample evaluation data group includes:

[0082] Get the verification condition parameters;

[0083] Obtain a calibration algorithm, and input the calibration condition parameters and the immune data sample in the sample evaluation data group into the calibration algorithm to obtain a calibration bias;

[0084] Obtaining a verification threshold and comparing the verification threshold with a verification bias; and when the verification bias is greater than the verification threshold, indicating that the corresponding sample evaluation data group passes the verification; otherwise, indicating that the corresponding sample evaluation data group fails the verification;

[0085] Count the percentage of sample evaluation data groups that have passed the verification in all sample evaluation data groups, and record it as the verification pass rate of the sample evaluation data group;

[0086] When performing verification processing on the immune data samples in each sample evaluation data group, the determined verification condition parameters are first obtained, and then the verification can be performed through the verification program. The verification algorithm here can be a pre-set verification function, and the obtained verification condition parameters and each immune data sample in the sample evaluation data group are input into the verification function together, and the verification bias is obtained by calculation. Among them, one of the preferred verification functions is expressed as: , where Indicates the calibration bias. Indicates immune data samples that rank low in the same group. Indicates the immune data sample that ranks high in the same group. It represents the time interval between the immune data samples in the same group, and then obtains the pre-set verification threshold, and compares this verification threshold with the obtained verification deviation. If the verification deviation is less than the verification threshold, it indicates that the corresponding sample evaluation data group has successfully passed the verification. Otherwise, it indicates that the sample evaluation data group has failed to pass the verification. Finally, statistics are performed on all sample evaluation data groups, and the specific proportion of sample evaluation data groups that have passed the verification in all sample evaluation data groups is calculated, and this proportion is recorded as the verification pass rate of the sample evaluation data group. The verification pass rate can not only reflect the stability or regularity of the immune data samples in the time series, but also can dynamically determine the length of the subsequent selected time period through the value of the verification pass rate.

[0087] Specifically, the step of determining the length of the selection period of the immune data sample according to the verification pass rate of each sample evaluation data group includes:

[0088] Obtain the verification pass rate and record it as the parameter to be evaluated;

[0089] Obtaining an evaluation interval, wherein multiple evaluation intervals are set, and each evaluation interval corresponds to an initial selection period;

[0090] Compare the parameter to be evaluated with the evaluation interval, match the corresponding initial selection period, and collect the number of immune data samples of the sample evaluation data group within the initial selection period, and record it as the selection condition parameter;

[0091] Obtaining an evaluation threshold and comparing the evaluation threshold with a selection condition parameter;

[0092] When the selection condition parameter is greater than the evaluation threshold, the initial selection period is directly determined as the selection period of the immune data sample. Otherwise, the length of the initial selection period is extended, and the selection condition parameters corresponding to the sample evaluation data group in the extended initial selection period are collected again until the selection condition parameter is greater than the evaluation threshold, and the extended initial selection period at this time is determined to be the length of the selection period of the immune data sample.

[0093] When determining the required selection period length, first collect the verification pass rate of each sample evaluation data group and record it as the parameter to be evaluated, and then obtain multiple evaluation intervals. It should be noted that the evaluation interval is pre-set, and each evaluation interval corresponds to a specific initial selection period. Then, compare the parameter to be evaluated with the multiple evaluation intervals one by one, and match the corresponding initial selection period through comparison. Collect the number of immune data samples of the sample evaluation data group within the initial selection period and record it as the selection condition parameter. Then introduce the pre-set evaluation threshold and compare and analyze the evaluation threshold with the selection condition parameter. If the selection condition parameter is greater than the evaluation threshold, the number of immune data samples of the sample evaluation data group is recorded as the selection condition parameter. If the evaluation threshold is exceeded, then this initial selection period can be directly determined as the selection period for the immune data sample. Conversely, if the selection condition parameter is less than the evaluation threshold, it is necessary to extend the length of the initial selection period and re-collect the selection condition parameters corresponding to the sample evaluation data group within the extended initial selection period. Repeat this process until the selection condition parameter is greater than the evaluation threshold. At this time, the extension can be stopped. The specific extension method can be extended according to a fixed ratio, etc., so as to ensure that there is a sufficient sample size for subsequent analysis, and the extended initial selection period at this time is determined as the final selection period length for the immune data sample, thereby providing corresponding data support for the determination of valid samples.

[0094] Furthermore, the patient's immune status is assessed through immune data samples, including:

[0095] After determining the valid samples, the valid samples are input into the evaluation function, and the output result of the evaluation function is recorded as the evaluation condition parameter;

[0096] Obtaining the state classification interval, adjusting the state classification interval with reference to the individualized fluctuation threshold, and then comparing it with the evaluation condition parameter;

[0097] When the evaluation condition parameter is in the adjusted status classification interval, it indicates that the patient's immune status is normal, and the patient's immune status is recorded as normal;

[0098] When the evaluation condition parameter exceeds the adjusted status classification interval, it indicates that the patient's immune status is abnormal, and the patient's immune status is recorded as abnormal.

[0099] In the above process, when evaluating the patient's immune status based on valid immune data samples, first obtain valid immune data samples and a pre-set evaluation function, and input the above-obtained valid immune data samples one by one into the evaluation function. Through the function's operation processing, the corresponding output results are generated, and these output results are recorded in detail as the basis for subsequent evaluation, that is, the evaluation condition parameters. The expression of the evaluation function is: , where Represents the evaluation condition parameters, Represents the valid immune data sample at the last node of the sample period, Indicates the required monitoring time, which is set according to the patient's review time. For example, the length of time between the patient's immune status being normal after the current examination and the next review. represents the length of the sample period, represents the number of valid sample condition parameters, and Represents the conditional parameters of adjacent valid samples, and obtains the pre-set status classification interval. The status classification interval is determined based on a large amount of clinical data and medical research, and can reflect the range of different immune states. Subsequently, the status classification interval is adjusted in combination with the individualized fluctuation threshold. The obtained evaluation condition parameters are compared with the status classification interval to determine whether the patient's immune status is normal. The real-time evaluation directly uses the valid immune data sample at the last node of the sample period as the evaluation condition parameter to compare with the status classification interval. The prognostic evaluation is based on the evaluation condition parameter calculated under the prognostic evaluation condition in the evaluation function as the benchmark for analysis. When the evaluation condition parameter is exactly within the status classification interval, it can be judged that the patient's immune status is within the normal range. At this time, the patient's immune status needs to be clearly recorded as normal for subsequent medical decision-making and health management. On the contrary, when the evaluation condition parameter exceeds the limit of the status classification interval, it indicates that the patient's immune status is abnormal and needs attention. At this time, the patient's immune status should be recorded as abnormal, and further appropriate medical measures should be taken for intervention and treatment. The status classification interval here.

[0100] S3. Under normal conditions and without postoperative complications, the corresponding immune data samples are collected and summarized as a normal state parameter set. If there are postoperative complications, it is re-marked as an abnormal state.

[0101] Under normal conditions, corresponding valid samples will be collected and summarized into a normal state parameter set. Specifically, when the patient's immune status is normal, corresponding valid samples will be collected and summarized into a normal state parameter set for doctors to refer to and analyze. Through the accumulation of a large amount of clinical data and immune data samples, the normal state parameter set can provide doctors with more accurate and comprehensive patient immune status information, and provide strong data support for subsequent medical decisions and health management.

[0102] S4. In abnormal conditions, the early warning mechanism is triggered and the immune deviation data is collected simultaneously. Combined with postoperative complications, the prognostic risk level is determined, including:

[0103] Obtaining the immune deviation value, where the immune deviation value is the difference between the evaluation condition parameter and the upper and lower limits of the adjusted status classification interval, and correcting the immune deviation value by combining the postoperative complications through weighted fusion;

[0104] Obtaining a grading interval, where multiple grading intervals are set, and each grading interval corresponds to a prognostic risk level;

[0105] Match the corrected immune deviation amount with the grading interval to obtain the corresponding prognostic risk level;

[0106] Among them, the higher the prognostic risk level, the further the patient's immune status deviates from the normal range.

[0107] To determine a patient's prognostic risk based on immune deviation, the immune deviation level must first be determined. This is calculated by taking the difference between the assessment parameter and the upper and lower limits of the state classification interval. If the assessment parameter exceeds the state classification interval, the difference is calculated with the upper limit of the state classification interval; otherwise, the difference is calculated with the lower limit of the state classification interval. Postoperative complications are then combined to create a severity score, with scores increasing in order from mild symptoms such as low-grade fever to toxic symptoms such as pneumonia to toxic symptoms such as sepsis. The number of postoperative complications is then recorded, and a weighted fusion is used to obtain a comprehensive complication score: C = S × a1 + N × a2, where C is the comprehensive complication score, S is the complication symptom score, and N is the number of occurrences. a1 and a2 are weights for the complication score and number of occurrences, respectively, which can be adjusted based on clinical importance, with a1 + a2 = 1. This formula is a conventional weighted fusion formula. The standard formula is referenced here for illustration purposes only. Subsequently, the comprehensive complication score and the immune deviation amount are weighted and fused again. The specific method refers to the above-mentioned weighted fusion formula to output the comprehensive risk score. Then, it is necessary to clarify the division of the grading intervals. The grading intervals are pre-set to multiple different ranges, and each grading interval uniquely corresponds to a specific prognostic risk level. The calculated comprehensive risk score is then matched one by one with the preset grading intervals to determine the prognostic risk level corresponding to the patient. In this process, the higher the value of the prognostic risk level, the more significant the deviation of the patient's immune status from the normal range, reflecting that the patient may face higher health risks.

[0108] See Figure 2 , a postoperative immune status monitoring and prognosis assessment system for thymoma patients, comprising:

[0109] The preoperative clustering module is used to cluster patients using a cluster analysis algorithm, construct a multi-level immune status feature cluster, and determine each patient's immune status category and individualized fluctuation threshold based on the feature cluster. Clustering includes preliminary clustering using a coarse-grained clustering algorithm and refined clustering using a density clustering algorithm.

[0110] The sample collection module is used to collect immune data samples of patients at consecutive time points after surgery;

[0111] The immune assessment module evaluates the patient's immune status using immune data samples. The assessment process determines whether the immune status is normal or abnormal based on individualized fluctuation thresholds.

[0112] The aggregation module collects corresponding immune data samples when the patient is in a normal state and there are no postoperative complications, and summarizes them into a normal state parameter set. If there are postoperative complications, the data is re-marked as an abnormal state.

[0113] The alarm module is used to trigger the early warning mechanism under abnormal conditions, and simultaneously collect the amount of immune deviation, and determine the prognostic risk level in combination with postoperative complications.

[0114] In the above system, the usage method thereof refers to the above-mentioned method for monitoring the immune status and evaluating the prognosis of thymoma patients after surgery, and will not be described in detail here.

[0115] See Figure 3 , an electronic device, the electronic device comprising:

[0116] at least one processor;

[0117] and a memory communicatively coupled to the at least one processor;

[0118] The memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can execute the above-mentioned method for monitoring the postoperative immune status and evaluating the prognosis of thymoma patients.

[0119] The processor of the electronic device can be a central processing unit (CPU), a graphics processing unit (GPU), or a digital signal processor (DSP), or can also be other general-purpose processors, digital signal processors, application-specific integrated circuits, field-programmable gate arrays or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The processor can implement or execute the methods, instructions, and programs disclosed herein. The processor can also be a microprocessor or any conventional processor. The memory of the electronic device can include a read-only memory (ROM), a programmable read-only memory (PROM), an electrically programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), flash memory, a hard disk, a CD-ROM, or other form of storage medium. The memory is used to store instructions, and the processor loads and executes the instructions to implement the above-mentioned method for monitoring the postoperative immune status and evaluating the prognosis of thymoma patients. The electronic device can also include an arithmetic unit, an input device, and an output device. The arithmetic unit can be an arithmetic logic unit (ALU) for performing various arithmetic and logical operations; an input device such as a keyboard, a mouse, or a touch screen for receiving user operation instructions and input data; and an output device such as a display or a printer for displaying operation results and related information.

[0120] It should be noted that, in this document, the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, apparatus, article, or method comprising a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, apparatus, article, or method. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, apparatus, article, or method comprising the element.

[0121] The foregoing is merely a preferred embodiment of the present invention. It should be noted that those skilled in the art may make various improvements and modifications without departing from the principles of the present invention, and such improvements and modifications are also within the scope of protection of the present invention. Structures, devices, and operating methods not specifically described or explained herein shall, unless otherwise specified or limited, be implemented in accordance with conventional means in the art.

Claims

1. A system for monitoring the immune status and evaluating the prognosis of postoperative thymoma patients, characterized by: include: The preoperative clustering module is used to cluster patients based on preoperative immune status data using a cluster analysis algorithm, construct a multi-level immune status feature cluster, and determine each patient's immune status category and individualized fluctuation threshold based on the feature cluster. Preoperative immune status data includes cell subset ratios. Clustering includes preliminary clustering using a coarse-grained clustering algorithm and refined clustering using a density clustering algorithm. The sample collection module is used to collect immune data samples of patients at consecutive time points after surgery; The immune assessment module evaluates the patient's immune status using immune data samples. The assessment process determines whether the immune status is normal or abnormal based on individualized fluctuation thresholds. The aggregation module collects corresponding immune data samples when the patient is in a normal state and there are no postoperative complications, and summarizes them into a normal state parameter set. If there are postoperative complications, the data is re-marked as an abnormal state. The alarm module is used to trigger the early warning mechanism in abnormal conditions, simultaneously collect immune deviation data, and determine the prognostic risk level based on postoperative complications; The method involves clustering patients based on preoperative immune status data using a cluster analysis algorithm to construct a multi-level immune status feature cluster, including: The proportion of T cell subsets was detected from the patient's preoperative blood samples and recorded as preoperative immune status data; Transfer learning is performed on preoperative immune status data through pre-trained immune feature encoders to generate high-order feature vectors. Dimensionality reduction tools are used to map high-order feature vectors to low-dimensional visualization space to obtain mapping results. A coarse-grained clustering algorithm was used to preliminarily group the mapping results and determine the distribution of the preset immune activity categories; Apply density clustering algorithm to refine the mapping results and set dynamic threshold parameters to identify small subgroups not covered by coarse-grained clustering algorithm; Based on the preliminary clustering results of the coarse-grained clustering algorithm and the tiny subgroups identified by the density clustering algorithm, a multi-level immune status characteristic distribution group was formed.

2. The postoperative immune status monitoring and prognosis evaluation system for thymoma patients according to claim 1, characterized in that: After collecting the immune data samples of the patient at consecutive time points after surgery, the immune data samples need to be preprocessed. The preprocessing method includes: The immune data samples are timestamped, and similar data at adjacent time points are merged through fluctuation parameter analysis to generate standardized time series parameters. Valid samples are screened from the standardized time series parameters, and interfering data are eliminated in combination with antibiotic usage markers.

3. The postoperative immune status monitoring and prognosis evaluation system for thymoma patients according to claim 2, characterized in that: The steps of obtaining the valid samples include: Samples are screened based on standardized timing parameters, and the screening results are recorded as verification parameters; According to the standardized time series parameters, the immune data samples are grouped in pairs according to the arrangement order to obtain multiple groups of sample evaluation data groups; Based on the verification parameters, the immune data samples in each sample evaluation data group are verified and the verification pass rate of the sample evaluation data group is determined; Determine the length of the selection period for the immune data sample based on the pass rate of each verification; The last node of the timestamp in the immune data sample is used as the reference node, and the reference node is reversely offset according to the length of the selected period to obtain the sample period; Immunity data samples covered within the sample period are recorded as valid samples, and samples outside the sample period are recorded as invalid samples.

4. The postoperative immune status monitoring and prognosis evaluation system for thymoma patients according to claim 3, characterized in that: The step of determining the verification pass rate of the sample evaluation data set includes: Get the verification condition parameters; Obtain a calibration algorithm, and input the calibration condition parameters and the immune data sample in the sample evaluation data group into the calibration algorithm to obtain a calibration bias; Obtaining a verification threshold and comparing the verification threshold with a verification bias; if the verification bias is less than the verification threshold, it indicates that the corresponding sample evaluation data group has passed the verification; otherwise, it indicates that the corresponding sample evaluation data group has failed the verification; The percentage of sample evaluation data groups that passed the verification in all sample evaluation data groups is counted and recorded as the verification pass rate of the sample evaluation data group.

5. The postoperative immune status monitoring and prognosis evaluation system for thymoma patients according to claim 3, characterized in that: The method of determining the length of the selection period of the immune data sample according to the verification pass rate of each sample evaluation data group includes: Obtain the verification pass rate and record it as the parameter to be evaluated; Obtaining an evaluation interval, wherein multiple evaluation intervals are set, and each evaluation interval corresponds to an initial selection period; Compare the parameter to be evaluated with the evaluation interval, match the corresponding initial selection period, and collect the number of immune data samples of the sample evaluation data group within the initial selection period, and record it as the selection condition parameter; Obtaining an evaluation threshold and comparing the evaluation threshold with a selection condition parameter; When the selection condition parameter is greater than the evaluation threshold, the initial selection period is directly determined as the selection period of the sample parameter. Otherwise, the length of the initial selection period is extended, and the selection condition parameters corresponding to the sample evaluation data group within the extended initial selection period are collected again until the selection condition parameter is greater than the evaluation threshold, and the extended initial selection period at this time is determined as the length of the selection period of the sample parameter.

6. The postoperative immune status monitoring and prognosis evaluation system for thymoma patients according to claim 1, characterized in that: The evaluation of the patient's immune status through the immune data sample includes: After determining the valid samples, the valid samples are input into the evaluation function, and the output result of the evaluation function is recorded as the evaluation condition parameter; Obtaining the state classification interval, adjusting the state classification interval with reference to the individualized fluctuation threshold, and then comparing it with the evaluation condition parameter; When the evaluation condition parameter is in the adjusted status classification interval, it indicates that the patient's immune status is normal, and the patient's immune status is recorded as normal; When the evaluation condition parameter exceeds the adjusted status classification interval, it indicates that the patient's immune status is abnormal, and the patient's immune status is recorded as abnormal.

7. The postoperative immune status monitoring and prognosis evaluation system for thymoma patients according to claim 1, characterized in that: In the abnormal state, the early warning mechanism is triggered, and the immune deviation amount is collected simultaneously. The prognostic risk level is determined in combination with postoperative complications, including: Obtaining the immune deviation value, where the immune deviation value is the difference between the evaluation condition parameter and the upper and lower limits of the adjusted status classification interval, and correcting the immune deviation value by combining the postoperative complications through weighted fusion; Obtaining a grading interval, where multiple grading intervals are set, and each grading interval corresponds to a prognostic risk level; Match the corrected immune deviation amount with the grading interval to obtain the corresponding prognostic risk level; Among them, the higher the prognostic risk level, the further the patient's immune status deviates from the normal range.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively coupled to the at least one processor; Wherein, the memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the functions of the postoperative immune status monitoring and prognosis evaluation system for thymoma patients according to any one of claims 1 to 7.

Images