A breast cancer stratified treatment strategy analysis method and system

Abstract

The application relates to the technical field of medical data analysis, and discloses a breast cancer hierarchical treatment strategy analysis method and system. The system comprises a multi-source data integration module, a risk stratification analysis module, a treatment strategy optimization module, an intelligent decision support module and a compliance supervision module. Multidimensional clinical data and biomarker data of breast cancer patients are collected, data preprocessing and standardization processing are carried out, the integrity and consistency of different source data are ensured, risk stratification processing is carried out based on the standardized patient data set, multilevel risk classification results are generated in real time, accurate risk identification is realized for different patients, data analysis errors are reduced, and the reliability of treatment strategy formulation is improved. The patient risk stratification data is matched and analyzed with a preset treatment strategy library, an optimal treatment strategy combination is searched by combining an artificial intelligence optimization algorithm, and the individuality and applicability of the treatment scheme are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical data analysis, in particular to a breast cancer stratified treatment strategy analysis method and system. BACKGROUND

[0002] Breast cancer, simply breast cancer, is a common malignant tumor in women, which is caused by uncontrolled proliferation of breast epithelial cells. High-risk factors for cancer include genetics and estradiol exposure, such as early menarche, late menopause, infertility, and chest radiation exposure during childhood. In addition to high-risk groups, late development, long lactation period, and early childbearing reduce the chances of breast cancer. Tumor types are divided into non-invasive cancer and invasive cancer and can develop to many organs in the body.

[0003] At present, due to the multi-source and heterogeneous characteristics of the clinical data and biomarker information of breast cancer patients, the conventional method usually relies on the experience of doctors and single-dimensional indicators when analyzing the stratified treatment strategy of breast cancer. It is difficult to standardize and integrate the multi-dimensional data of patients and dynamically stratify the risk. If the analysis is based on incomplete or inaccurate data, it may cause a large deviation in the risk classification of patients, and cannot guarantee the accuracy of the subsequent treatment strategy.

[0004] Therefore, the present application provides a breast cancer stratified treatment strategy analysis method and system to solve the above problems. SUMMARY

[0005] In view of the deficiencies of the prior art, the present application provides a breast cancer stratified treatment strategy analysis method and system to solve the problem of deviation in the risk classification of patients caused by incomplete or inaccurate data analysis in the background art, which cannot guarantee the accuracy of the subsequent treatment strategy.

[0006] To achieve the above purpose, the present application provides the following technical solution: a breast cancer stratified treatment strategy analysis method and system, the method comprising the following steps: S1, collecting multi-dimensional clinical data and biomarker data of breast cancer patients, including patient basic information, pathological parameters, genomics data, imaging features and treatment history information, generating a patient original data set; S2, data preprocessing and standardization processing based on the patient original data set, including missing value filling, outlier correction, data normalization and feature scaling, generating a standardized patient data set; S3, breast cancer risk stratification processing according to the standardized patient data set, using a machine learning clustering algorithm to classify patients at multiple levels of risk, generating patient risk stratification data, wherein the risk levels include low risk, medium risk and high risk categories; S4, based on the patient risk stratification data and the preset treatment strategy library, a matching analysis is performed, an optimal treatment strategy combination is searched by combining an artificial intelligence optimization algorithm, and preliminary stratified treatment strategy recommendation data are generated; S5, treatment strategy effectiveness simulation verification processing is performed, treatment effects under different strategies are simulated by constructing a virtual patient model, and treatment strategy verification data are generated; S6, consistency evaluation is performed according to the treatment strategy verification data and a clinical guideline benchmark, and a strategy optimization adjustment instruction is generated; S7, based on the strategy optimization adjustment instruction, the preliminary stratified treatment strategy recommendation data are dynamically corrected, and a final breast cancer stratified treatment strategy analysis report is output.

[0007] Preferably, the breast cancer patient multi-dimensional clinical data and biomarker data collected in S1 include the following steps: S11, patient basic information is extracted through an electronic health record system, including age, gender, family history and complications, and patient clinical basic data are generated; S12, pathological laboratory system is integrated to obtain tumor grade, hormone receptor status, HER2 expression level and Ki-67 index, and pathological parameter data are generated; S13, genomic sequencing platform is used to collect gene mutation, expression profile and multi-omics data, and a genomics data set is generated; S14, breast X-ray photography, ultrasound and MRI image features are extracted through a medical image archive system, including tumor size, shape and texture features, and imaging feature data are generated; S15, patient past surgery, chemotherapy, radiotherapy and targeted therapy history are obtained from a treatment record library, and a treatment history information set is generated; S16, the patient clinical basic data, pathological parameter data, genomics data set, imaging feature data and treatment history information set are fused, the patient original data set is constructed, and data encryption technology is used to ensure privacy and security.

[0008] Preferably, the data preprocessing and standardization processing in S2 include the following steps: S21, the patient original data set is obtained, missing values and abnormal values in the data are identified, interpolation method is used for missing value filling, and abnormal values are removed based on box plot rule; S22, minimum-maximum normalization processing is performed on continuous features, and one-hot encoding is performed on category features, so as to realize data scale unification; S23, a feature selection algorithm is applied, including chi-square test, to remove redundant features and generate a high-dimensional feature subset; S24. Perform feature dimensionality reduction through principal component analysis to reduce computational complexity, generate the standardized patient dataset, and store it in a cloud database for subsequent analysis.

[0009] Preferably, the breast cancer risk stratification process in S3 includes the following steps: S31. Load the standardized patient dataset and initialize the clustering algorithm parameters, including the number of clusters K and the maximum number of iterations; S32. Perform clustering algorithm calculation to calculate the cosine similarity between patient samples and assign the samples to the nearest centroid. S33. Iterate and update the centroid position until convergence, and output the initial cluster labels; S34. Use the silhouette coefficient to evaluate the clustering quality. If the quality is below the threshold, adjust the parameters and re-cluster. S35. Divide risk levels according to clustering results: Low-risk strata correspond to good prognostic characteristics: hormone receptor positivity and low proliferation index; The medium-risk level corresponds to a medium probability of recurrence. High-risk strata correspond to triple negative results and high metastasis potential; S36. Generate the patient risk stratification data and display the stratification distribution in the form of a visual chart to assist clinicians in interpretation.

[0010] Preferably, the matching analysis in S4 includes the following steps: S41. Construct a preset treatment strategy library, which stores various treatment strategy templates, including surgical methods, chemotherapy regimens, radiotherapy plans, endocrine therapy and immunotherapy options, with each strategy associated with efficacy indicators and side effect scores; S42. Match the patient risk stratification data with the treatment strategy library, and screen candidate strategies based on risk stratification; S43. Using artificial intelligence optimization algorithms, with the goal of prolonging patient survival and improving quality of life, search for the optimal combination of strategies; S44. By balancing efficacy and toxicity through multi-objective optimization, a Pareto optimal solution set is generated. S45. Based on the knowledge of clinical experts, the strategies are ranked, and the preliminary stratified treatment strategy recommendation data is output, with confidence scores marked.

[0011] Preferably, the generation of treatment strategy verification data in step S5 includes the following steps: S51. A virtual patient model is constructed based on a physiological pharmacokinetic model, with model parameters derived from individual patient characteristics; S52. Load preliminary stratified treatment strategy recommendation data and simulate tumor volume changes, drug concentration dynamics and adverse events after strategy implementation; S53. Monte Carlo simulation was used to handle random factors and the simulation output indicators were calculated, including progression-free survival, overall survival and toxicity incidence. S55. Compare the simulation results with historical real-world data to verify the model's calibration and generate validation data for the treatment strategy.

[0012] Preferably, the consistency assessment in S6 includes the following steps: S61. Obtain international clinical guidelines and benchmarks: NCCN and ESMO guidelines, and extract standard treatment recommendations; S62. Compare the validation data of the treatment strategy with the guideline benchmarks and calculate the compliance index; S63. Use the statistical test method t-test to assess the significance of the difference; S64. When the compliance rate is lower than the preset threshold, generate strategy adjustment suggestions, including enhanced monitoring and optimization of combination therapy; S65. Output strategy optimization and adjustment instructions, and record the reasons for deviations for learning and optimization.

[0013] Preferably, the output of the final breast cancer stratified treatment strategy analysis report in step S7 includes the following steps: S71. Parse the strategy optimization and adjustment instructions to identify the strategy parameters that need to be corrected; S72. Apply adaptive control algorithms to dynamically adjust treatment dosage and timing, and update in real time based on feedback data; S73. Integrate real-time monitoring data: circulating tumor DNA, imaging review results, and further optimize the strategy; S74. Generate the final breast cancer stratification treatment strategy analysis report, which includes the stratification basis, recommended strategies, simulation validation results, and clinical implementation suggestions; S75. Push reports to clinical terminals through the medical decision support system, and support interactive query and revision.

[0014] Preferably, the system includes: The multi-source data integration module acquires and processes multi-dimensional patient information through data acquisition and standardization units, generating a standardized patient dataset. The risk stratification analysis module receives the standardized patient dataset, classifies patients through a clustering calculation unit, and outputs patient risk stratification data through a stratification quality assessment unit. The treatment strategy optimization module receives the patient risk stratification data, generates a preliminary strategy through the strategy matching and simulation unit, optimizes it using the dynamic verification and adjustment unit, and outputs a final breast cancer stratification treatment strategy analysis report.

[0015] Preferably, the system further includes: The intelligent decision support module receives user preferences and clinical constraints from external input, processes them through the constraint parsing unit and the reinforcement learning optimization unit, generates personalized strategy adjustment instructions, and feeds them back to the treatment strategy optimization module. The compliance and oversight module ensures that the strategy complies with regulations through the ethics rule verification unit and records the entire system operation process through the audit log generation unit, outputting a compliance audit report.

[0016] Compared with existing technologies, the present invention provides a method and system for analyzing stratified treatment strategies for breast cancer, which has the following beneficial effects: 1. In this invention, multidimensional clinical data and biomarker data of breast cancer patients are collected, and data preprocessing and standardization are performed to ensure the integrity and consistency of data from different sources. At the same time, risk stratification is performed based on the standardized patient dataset to generate multi-level risk classification results in real time, thereby achieving accurate risk identification for different patients, reducing data analysis errors, and improving the reliability of treatment strategy formulation.

[0017] 2. In this invention, patient risk stratification data is matched and analyzed with a preset treatment strategy library. The optimal combination of treatment strategies is searched using artificial intelligence optimization algorithms. Preliminary stratified treatment strategy recommendation data is generated in real time. Based on simulation verification and consistency assessment, the strategy is dynamically corrected and treatment parameters are automatically adjusted to ensure the conformity of the recommended strategies with clinical guidelines and improve the personalization and applicability of the treatment plan.

[0018] 3. In this invention, a virtual patient model is constructed to simulate the treatment effect and verify the effectiveness of the strategy. Combined with the intelligent decision support module and the compliance supervision module, the output is optimized in real time to achieve multi-dimensional data integration and automated analysis, reduce the uncertainty of manual intervention, and improve the accuracy and overall efficiency of stratified treatment strategy analysis. Attached Figure Description

[0019] Fig. 1 This is a flowchart of a method for analyzing a stratified treatment strategy for breast cancer according to the present invention; Fig. 2 This is a schematic diagram of the architecture of a breast cancer stratified treatment strategy analysis system according to 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 skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] For specific implementation examples, please refer to: Figs. 1-2A method and system for analyzing stratified treatment strategies for breast cancer, the method comprising the following steps: S1. Collect multidimensional clinical data and biomarker data of breast cancer patients, including basic patient information, pathological parameters, genomic data, imaging features and treatment history information, and generate the patient's original dataset; S2. Based on the original patient dataset, perform data preprocessing and standardization, including missing value imputation, outlier correction, data normalization, and feature scaling, to generate a standardized patient dataset. S3. Based on the standardized patient dataset, perform breast cancer risk stratification. Use machine learning clustering algorithms to classify patients into multi-level risk categories and generate patient risk stratification data, where the risk levels include low risk, medium risk and high risk categories. S4. Based on the patient risk stratification data and the preset treatment strategy library, a matching analysis is performed, and the optimal combination of treatment strategies is searched using artificial intelligence optimization algorithms to generate preliminary stratified treatment strategy recommendation data. S5. Conduct treatment strategy effectiveness simulation and verification processing. Simulate the treatment effect under different strategies by constructing a virtual patient model and generate treatment strategy verification data. S6. Based on the consistency assessment between the treatment strategy validation data and clinical guideline benchmarks, generate strategy optimization and adjustment instructions; S7. Based on the strategy optimization and adjustment instructions, dynamically correct the preliminary stratified treatment strategy recommendation data and output the final breast cancer stratified treatment strategy analysis report.

[0022] S1 involves collecting multidimensional clinical and biomarker data from breast cancer patients, including the following steps: S11. Extract basic patient information, including age, gender, family history and complications, through the electronic health record system, and generate basic clinical data of the patient. S12. Integrate the pathology laboratory system to obtain tumor grade, hormone receptor status, HER2 expression level and Ki-67 index, and generate pathology parameter data; S13. Collect gene mutation, expression profile and multi-omics data using a genome sequencing platform to generate a genomics dataset; S14. Extract mammography, ultrasound and MRI image features through the medical image archiving system, including tumor size, shape and texture features, and generate imaging feature data; S15. Obtain the patient's past surgery, chemotherapy, radiotherapy and targeted therapy history from the treatment record database, and generate a treatment history information set; S16. Multi-source data fusion is performed on the patient's basic clinical data, pathological parameter data, genomic dataset, imaging feature data and treatment history information set to construct the patient's original dataset, and data encryption technology is used to ensure privacy and security. The implementation of multi-source data fusion first connects heterogeneous data sources through a data integration platform, and then uses an ontology-based semantic alignment method to map data from different sources into a unified knowledge graph to eliminate pattern conflicts. The data encryption technology employs the Advanced Encryption Standard (AES) combined with public key infrastructure. It uses transport layer security protocols during data transmission and implements field-level encryption during storage. Key management is based on a hardware security module, ensuring data privacy and compliance. The encryption process can be expressed as the formula: encryption: ; Deciphering: ; in For encrypted data, Plain text data For encryption keys; Data preprocessing and standardization in S2 include the following steps: S21. Obtain the patient's original dataset, identify missing and outlier values ​​in the data, use interpolation to fill in missing values, and remove outliers based on box plot rules. For time series data, linear interpolation is used, and the formula is: ; in For the missing values ​​to be estimated, This indicates the location of the data point containing the missing value. The value of the known data point preceding the location of the missing value. This is the position of the previous known data point. The value of the one known data point following the location of the missing value. This indicates the position of the next known data point. For outliers, remove them based on the box plot rule: calculate the quartiles of the data, Q1 being the 25th percentile and Q3 the 75th percentile, and calculate the interquartile range: IQR = Q3. Q1, define an outlier as a value lower than Q1. Data points with a value of 1.5×IQR or higher than Q3+1.5×IQR are directly discarded and corrected to boundary values. This step ensures data quality and reduces the impact of noise. S22. Perform min-max normalization on continuous features and one-hot encoding on categorical features to achieve data scale uniformity. The minimum-maximum normalization formula is: ; in These are the original eigenvalues. These are the normalized eigenvalues. For feature vectors; One-hot encoding is performed on categorical features, converting each category into a binary vector. When a category has n possible values, an n-dimensional vector is generated, with the corresponding category position set to 1 and the rest set to 0. This step achieves data scaling uniformity, which facilitates subsequent processing by machine learning algorithms. S23. Apply feature selection algorithms, including chi-square test, to remove redundant features and generate high-dimensional feature subsets; The chi-square test is used to assess the independence of the characteristic from the risk level of the target variable, and the chi-square statistic is calculated: ; in This is the chi-square statistic. To observe the frequency, To determine the expected frequency, features with p-values ​​below a threshold of 0.05 are selected as significant features, generating a high-dimensional feature subset. S24. Perform feature dimensionality reduction through principal component analysis to reduce computational complexity, generate a standardized patient dataset, and store it in a cloud database for subsequent analysis. Feature dimensionality reduction is performed using principal component analysis. First, the covariance matrix of the standardized dataset is calculated. Then, the eigenvalues ​​and eigenvectors are solved. The first k principal components are selected such that the cumulative variance contribution exceeds 85%. The projection formula is: ; in The original feature matrix, The eigenvector matrix, This is the dimensionality-reduced dataset.

[0023] S3 breast cancer risk stratification includes the following steps: S31. Load the standardized patient dataset and initialize the clustering algorithm parameters, including the number of clusters K and the maximum number of iterations; For K-means clustering, the number of clusters K is determined by the elbow rule: calculate the sum of squares within clusters (WSS) for different K values, and select the K value corresponding to the inflection point of WSS descent; set the maximum number of iterations to 100-500 to ensure convergence; after parameter initialization, pass the clustering algorithm framework Scikit-learn. S32. Perform clustering algorithm calculation to calculate the cosine similarity between patient samples and assign the samples to the nearest centroid. ; Among them A and Given the sample feature vector; based on the similarity matrix, assign each sample to the nearest centroid, and iteratively update the centroid position; S33. Iterate and update the centroid position until convergence, and output the initial cluster labels; S34. Use the silhouette coefficient to evaluate the clustering quality. If the quality is below the threshold, adjust the parameters and re-cluster. For each sample i, calculate the silhouette coefficient: ; in For the sample The profile coefficient, For the sample The average distance to other points in the same cluster. For the sample The average distance to the nearest other cluster; S35. Divide risk levels according to clustering results: Low-risk strata correspond to good prognostic characteristics: hormone receptor positivity and low proliferation index; The medium-risk level corresponds to a medium probability of recurrence. High-risk strata correspond to triple negative results and high metastasis potential; S36. Generate patient risk stratification data and display the stratification distribution in the form of visual charts to assist clinicians in interpretation.

[0024] Matching analysis in S4 includes the following steps: S41. Construct a preset treatment strategy library, which stores various treatment strategy templates, including surgical methods, chemotherapy regimens, radiotherapy plans, endocrine therapy and immunotherapy options, with each strategy associated with efficacy indicators and side effect scores; S42. Match patient risk stratification data with the treatment strategy library and screen candidate strategies based on risk stratification. S43. Using artificial intelligence optimization algorithms, with the goal of prolonging patient survival and improving quality of life, search for the optimal combination of strategies; Artificial intelligence optimization algorithm: Genetic algorithm, used to search for the optimal treatment strategy, implementation steps: Initialize the population, with each individual representing a strategy combination; calculate the fitness function; perform selection, crossover, and mutation operations: selection is based on roulette wheel selection, crossover uses single-point crossover, and mutation randomly modifies the strategy parameters; iteratively update the population until convergence; the fitness function can be defined as: in For the fitness function value, , These are the weighting coefficients; S44. By balancing efficacy and toxicity through multi-objective optimization, a Pareto optimal solution set is generated. When implementing multi-objective optimization to balance efficacy and toxicity, the NSGA-II algorithm is used to generate a Pareto optimal solution set: the objective functions are defined as efficacy indicators and toxicity indicators, the diversity of the solution set is maintained by non-dominated sorting and crowding calculation, and finally a set of balanced solutions is output for clinical selection; S45. Based on the knowledge of clinical experts, the strategies are ranked, and preliminary stratified treatment strategy recommendation data is output, with confidence scores marked.

[0025] The process of generating treatment strategy validation data in S5 includes the following steps: S51. A virtual patient model is constructed based on a physiological pharmacokinetic model, with model parameters derived from individual patient characteristics; Using differential equations to simulate the compartmentalization of drugs in the body: ; in This represents the rate of change in drug concentration. Drug concentration, For time variables, To eliminate the rate constant, the model is integrated into the simulation platform to generate virtual patient responses. S52. Load preliminary stratified treatment strategy recommendation data and simulate tumor volume changes, drug concentration dynamics and adverse events after strategy implementation; S53. Monte Carlo simulation was used to handle random factors and the simulation output indicators were calculated, including progression-free survival, overall survival and toxicity incidence. By simulating random events under the treatment strategy through random sampling, the output indicators were calculated: progression-free survival based on the tumor growth model, overall survival based on survival analysis, and toxicity incidence based on probability distribution. The simulation was repeated more than 1000 times to generate the probability distribution of the indicators and verify the robustness of the strategy. S55. Compare the simulation results with historical real-world data to verify the model's calibration and generate validation data for the treatment strategy.

[0026] The consistency assessment in S6 includes the following steps: S61. Obtain international clinical guidelines and benchmarks: NCCN and ESMO guidelines, and extract standard treatment recommendations; S62. Compare the treatment strategy validation data with the guideline benchmarks and calculate the compliance index; ; in For compliance indicators, For the recommendation policy vector, This serves as the reference vector for the guide. S63. Use the statistical test method t-test to assess the significance of the difference; The formula for the independent samples t-test is: ; in Indicates the degree of difference between the means of two groups of samples. This is the mean of the first group of samples. This is the mean of the second group of samples. The variance of the first sample group. The variance of the second group of samples. The size of the first group of samples, The size of the second group of samples; S64. When the compliance rate is lower than the preset threshold, generate strategy adjustment suggestions, including enhanced monitoring and optimization of combination therapy; S65. Output strategy optimization and adjustment instructions, and record the reasons for deviations for learning and optimization.

[0027] The final breast cancer stratified treatment strategy analysis report is output in S7, including the following steps: S71. Parse the strategy optimization and adjustment instructions and identify the strategy parameters that need to be corrected; S72. Apply adaptive control algorithms to dynamically adjust treatment dosage and timing, and update in real time based on feedback data; A patient state-space model is established, with dose and time sequence as control inputs. In each control cycle, the future state is predicted based on real-time monitoring data, and the minimum objective function of the optimization problem is solved. The strategy is adjusted and implemented, with feedback used to update the model. Formulaically, the control law can be expressed as: ; in To control the input, The deviation between the target value and the actual value. For adaptive parameters; S73. Integrate real-time monitoring data: circulating tumor DNA, imaging review results, and further optimize the strategy; S74. Generate the final breast cancer stratification treatment strategy analysis report, which includes the stratification basis, recommended strategies, simulation validation results, and clinical implementation suggestions; S75. Push reports to clinical terminals through the medical decision support system, and support interactive query and revision.

[0028] The system includes: The multi-source data integration module acquires and processes multi-dimensional patient information through data acquisition and standardization units, generating a standardized patient dataset. The risk stratification analysis module receives a standardized patient dataset, classifies patients through a clustering calculation unit, and outputs patient risk stratification data through a stratification quality assessment unit. The treatment strategy optimization module receives patient risk stratification data, generates preliminary strategies through strategy matching and simulation units, optimizes them using dynamic verification and adjustment units, and outputs a final breast cancer stratified treatment strategy analysis report.

[0029] The system also includes: The intelligent decision support module receives user preferences and clinical constraints from external input, processes them through the constraint parsing unit and the reinforcement learning optimization unit, generates personalized strategy adjustment instructions, and feeds them back to the treatment strategy optimization module. The compliance and oversight module ensures that the strategy complies with regulations through the ethics rule verification unit and records the entire system operation process through the audit log generation unit, outputting a compliance audit report.

[0030] The operation steps of this method and system are as follows: Step 1: Multi-source data acquisition and standardization processing: Basic patient information is extracted from the electronic health record system, pathological parameter data is obtained from the pathology laboratory system, genomic datasets are collected using a genome sequencing platform, imaging feature data is extracted from the medical imaging archive system, and treatment history information is obtained from the treatment record database. These multi-source data are then fused to generate the original patient dataset. Subsequent data preprocessing and standardization processes, including missing value imputation, outlier correction, data normalization, and feature scaling, result in a standardized patient dataset.

[0031] Step 2: Breast Cancer Risk Stratification: Based on a standardized patient dataset, a machine learning clustering algorithm was used for multi-level risk classification. The clustering algorithm parameters were initialized, cosine similarity between patient samples was calculated, and the centroid positions were iteratively updated until convergence, generating initial cluster labels. The silhouette coefficient was used to evaluate clustering quality, and the clustering results were used to divide the data into three levels: low-risk, medium-risk, and high-risk. The low-risk level corresponds to good prognostic characteristics, the medium-risk level corresponds to a moderate recurrence probability, and the high-risk level corresponds to triple-negative status and high metastatic potential.

[0032] Step 3: Matching and Optimizing Treatment Strategies Patient risk stratification data is matched and analyzed against a pre-defined treatment strategy library to screen candidate strategies based on risk levels. An artificial intelligence optimization algorithm is used to search for the optimal combination of treatment strategies, balancing efficacy and toxicity through multi-objective optimization to generate a Pareto optimal solution set. Strategies are ranked in conjunction with clinical expert knowledge, outputting preliminary stratified treatment strategy recommendations with confidence scores.

[0033] Step 4: Strategy Verification and Consistency Assessment A virtual patient model was constructed based on a physiological pharmacokinetic model to simulate tumor volume changes, drug concentration dynamics, and adverse events after the implementation of a treatment strategy. Monte Carlo simulation was used to handle random factors, and progression-free survival, overall survival, and toxicity incidence were calculated. The simulation results were compared with historical real-world data to verify the model's calibration and generate treatment strategy validation data. Subsequently, a consistency assessment was conducted with international clinical guidelines; when the compliance rate fell below a preset threshold, strategy adjustment recommendations were generated.

[0034] Step 5: Dynamic Correction and Report Generation Based on strategy optimization and adjustment instructions, the treatment dosage and timing are dynamically adjusted through an adaptive control algorithm. Real-time monitoring data, including circulating tumor DNA and imaging review results, are integrated for further strategy optimization. Finally, a breast cancer stratified treatment strategy analysis report is generated, including stratification criteria, recommended strategies, simulation validation results, and clinical implementation suggestions, and is pushed to clinical terminals through a medical decision support system.

[0035] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0036] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for analyzing stratified treatment strategies for breast cancer, characterized in that: The method includes the following steps: S1. Collect multidimensional clinical data and biomarker data of breast cancer patients, including basic patient information, pathological parameters, genomic data, imaging features and treatment history information, and generate the patient's original dataset; S2. Based on the original patient dataset, perform data preprocessing and standardization, including missing value imputation, outlier correction, data normalization, and feature scaling, to generate a standardized patient dataset. S3. Based on the standardized patient dataset, perform breast cancer risk stratification processing, and use machine learning clustering algorithm to classify patients into multi-level risk categories to generate patient risk stratification data, wherein the risk levels include low risk, medium risk and high risk categories. S4. Based on the patient risk stratification data and the preset treatment strategy library, a matching analysis is performed, and the optimal combination of treatment strategies is searched using artificial intelligence optimization algorithms to generate preliminary stratified treatment strategy recommendation data. S5. Conduct treatment strategy effectiveness simulation and verification processing. Simulate the treatment effect under different strategies by constructing a virtual patient model and generate treatment strategy verification data. S6. Based on the validation data of the treatment strategy, conduct a consistency assessment with clinical guideline benchmarks and generate strategy optimization and adjustment instructions; S7. Based on the strategy optimization and adjustment instructions, dynamically correct the preliminary stratified treatment strategy recommendation data and output the final breast cancer stratified treatment strategy analysis report.

2. The method for analyzing a stratified treatment strategy for breast cancer according to claim 1, characterized in that: The S1 step involves collecting multidimensional clinical data and biomarker data from breast cancer patients, including the following steps: S11. Extract basic patient information, including age, gender, family history and complications, through the electronic health record system, and generate basic clinical data of the patient. S12. Integrate the pathology laboratory system to obtain tumor grade, hormone receptor status, HER2 expression level and Ki-67 index, and generate pathology parameter data; S13. Collect gene mutation, expression profile and multi-omics data using a genome sequencing platform to generate a genomics dataset; S14. Extract mammography, ultrasound and MRI image features through the medical image archiving system, including tumor size, shape and texture features, and generate imaging feature data; S15. Obtain the patient's past surgery, chemotherapy, radiotherapy and targeted therapy history from the treatment record database, and generate a treatment history information set; S16. The patient's basic clinical data, pathological parameter data, genomic dataset, imaging feature data, and treatment history information are fused from multiple sources to construct the patient's original dataset, and data encryption technology is used to ensure privacy and security.

3. The method for analyzing a stratified treatment strategy for breast cancer according to claim 2, characterized in that: The data preprocessing and standardization process in S2 includes the following steps: S21. Obtain the original patient dataset, identify missing and outlier values ​​in the data, fill in missing values ​​using interpolation, and remove outliers based on box plot rules. S22. Perform min-max normalization on continuous features and one-hot encoding on categorical features to achieve data scale uniformity. S23. Apply feature selection algorithms, including chi-square test, to remove redundant features and generate high-dimensional feature subsets; S24. Perform feature dimensionality reduction through principal component analysis to reduce computational complexity, generate the standardized patient dataset, and store it in a cloud database for subsequent analysis.

4. The method for analyzing a stratified treatment strategy for breast cancer according to claim 3, characterized in that: The breast cancer risk stratification process in S3 includes the following steps: S31. Load the standardized patient dataset and initialize the clustering algorithm parameters, including the number of clusters K and the maximum number of iterations; S32. Perform clustering algorithm calculation to calculate the cosine similarity between patient samples and assign the samples to the nearest centroid. S33. Iterate and update the centroid position until convergence, and output the initial cluster labels; S34. Use the silhouette coefficient to evaluate the clustering quality. If the quality is below the threshold, adjust the parameters and re-cluster. S35. Divide risk levels according to clustering results: Low-risk strata correspond to good prognostic characteristics: hormone receptor positivity and low proliferation index; The medium-risk level corresponds to a medium probability of recurrence. High-risk strata correspond to triple negative results and high metastasis potential; S36. Generate the patient risk stratification data and display the stratification distribution in the form of a visual chart to assist clinicians in interpretation.

5. The method for analyzing a stratified treatment strategy for breast cancer according to claim 4, characterized in that: The matching analysis in S4 includes the following steps: S41. Construct a preset treatment strategy library, which stores various treatment strategy templates, including surgical methods, chemotherapy regimens, radiotherapy plans, endocrine therapy and immunotherapy options, with each strategy associated with efficacy indicators and side effect scores; S42. Match the patient risk stratification data with the treatment strategy library, and screen candidate strategies based on risk stratification; S43. Using artificial intelligence optimization algorithms, with the goal of prolonging patient survival and improving quality of life, search for the optimal combination of strategies; S44. By balancing efficacy and toxicity through multi-objective optimization, a Pareto optimal solution set is generated. S45. Based on the knowledge of clinical experts, the strategies are ranked, and the preliminary stratified treatment strategy recommendation data is output, with confidence scores marked.

6. The method for analyzing a stratified treatment strategy for breast cancer according to claim 5, characterized in that: The generation of treatment strategy verification data in S5 includes the following steps: S51. A virtual patient model is constructed based on a physiological pharmacokinetic model, with model parameters derived from individual patient characteristics; S52. Load preliminary stratified treatment strategy recommendation data and simulate tumor volume changes, drug concentration dynamics and adverse events after strategy implementation; S53. Monte Carlo simulation was used to handle random factors and the simulation output indicators were calculated, including progression-free survival, overall survival and toxicity incidence. S55. Compare the simulation results with historical real-world data to verify the model's calibration and generate validation data for the treatment strategy.

7. The method for analyzing a stratified treatment strategy for breast cancer according to claim 6, characterized in that: The consistency assessment in S6 includes the following steps: S61. Obtain international clinical guidelines and benchmarks: NCCN and ESMO guidelines, and extract standard treatment recommendations; S62. Compare the validation data of the treatment strategy with the guideline benchmarks and calculate the compliance index; S63. Use the statistical test method t-test to assess the significance of the difference; S64. When the compliance rate is lower than the preset threshold, generate strategy adjustment suggestions, including enhanced monitoring and optimization of combination therapy; S65. Output strategy optimization and adjustment instructions, and record the reasons for deviations for learning and optimization.

8. The method for analyzing a stratified treatment strategy for breast cancer according to claim 7, characterized in that: The final breast cancer stratified treatment strategy analysis report output in S7 includes the following steps: S71. Parse the strategy optimization and adjustment instructions to identify the strategy parameters that need to be corrected; S72. Apply adaptive control algorithms to dynamically adjust treatment dosage and timing, and update in real time based on feedback data; S73. Integrate real-time monitoring data: circulating tumor DNA, imaging review results, and further optimize the strategy; S74. Generate the final breast cancer stratification treatment strategy analysis report, which includes the stratification basis, recommended strategies, simulation validation results, and clinical implementation suggestions; S75. Push reports to clinical terminals through the medical decision support system, and support interactive query and revision.

9. A breast cancer stratified treatment strategy analysis system, used to implement the breast cancer stratified treatment strategy analysis method according to any one of claims 1-8, characterized in that: The system includes: The multi-source data integration module acquires and processes multi-dimensional patient information through data acquisition and standardization units, generating a standardized patient dataset. The risk stratification analysis module receives the standardized patient dataset, classifies patients through a clustering calculation unit, and outputs patient risk stratification data through a stratification quality assessment unit. The treatment strategy optimization module receives the patient risk stratification data, generates a preliminary strategy through the strategy matching and simulation unit, optimizes it using the dynamic verification and adjustment unit, and outputs a final breast cancer stratification treatment strategy analysis report.

10. A breast cancer stratified treatment strategy analysis system according to claim 9, characterized in that: The system also includes: The intelligent decision support module receives user preferences and clinical constraints from external input, processes them through the constraint parsing unit and the reinforcement learning optimization unit, generates personalized strategy adjustment instructions, and feeds them back to the treatment strategy optimization module. The compliance and oversight module ensures that the strategy complies with regulations through the ethics rule verification unit and records the entire system operation process through the audit log generation unit, outputting a compliance audit report.

Images