Tumor recurrence risk prediction method and system based on electronic medical record data

By constructing a tumor recurrence risk level mapping model based on graph neural network and integrating electronic medical record data, individualized evaluation and dynamic intervention of tumor recurrence risk are solved, and the problem of insufficient data integration and prediction accuracy in the existing technology is improved, and the accuracy of tumor recurrence risk prediction and clinical application efficiency are improved.

CN120544908AActive Publication Date: 2025-08-26CHENGDU MILITARY GENERAL HOSPITAL OF PLA

Patent Information

Application Number
CN202511030178.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-07-25
Publication Date
2025-08-26
Estimated Expiration
2045-07-25

AI Technical Summary

Technical Problem

The existing tumor recurrence risk prediction cannot effectively integrate structured, semi-structured and unstructured data in electronic medical records, resulting in the failure to effectively integrate the predicted results with clinical intervention operations, and staying in the retrospective research stage.

Method used

By collecting and preprocessing electronic medical record data of patients of multiple age groups, classification and mapping, a tumor recurrence risk level mapping model is constructed based on graph neural network, and dynamic intervention is carried out in combination with tumor marker parameters to form closed-loop management.

Benefits of technology

It realizes individualized recurrence risk assessment, reduces subjective judgment errors, improves the identification accuracy of high-risk patients, reduces the risk of excessive medical treatment, and improves clinical efficiency and prediction accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120544908A_ABST
    Figure CN120544908A_ABST
Patent Text Reader

Abstract

The invention discloses a tumor recurrence risk prediction method and system based on electronic medical record data, and relates to the field of electronic medical record data processing and analys.The historical electronic medical record data are processed in a structured mode, a recurrence risk mapping model is established, individualized recurrence risk assessment can be achieved based on multi-dimensional features, and the tumor recurrence risk prediction accuracy is improved. The subjective judgment error is obviously reduced; a risk level layering mechanism can automatically distinguish patients needing emergency intervention and conventional monitoring, excessive medical treatment or delayed treatment is avoided, and the method is particularly suitable for chronic diseases such as tumors needing long-term management; through periodic marker measurement and risk level feedback, closed-loop management of evaluation-intervention-re-evaluation is formed, and the requirement for continuous optimization of clinical diagnosis and treatment is met; in addition, through preprocessing, the preprocessed multi-dimensional features can be directly called subsequently, and repeated data cleaning work is avoided.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the field of electronic medical record data processing and analysis, and more specifically, relates to a method and system for predicting tumor recurrence risk based on electronic medical record data. Background Art

[0002] Existing tumor recurrence risk prediction cannot effectively integrate structured test indicators, semi-structured text descriptions (such as medical records) and unstructured imaging data in electronic medical records to construct a unified feature space for cross-modal correlation analysis; in addition, existing prediction models do not integrate prediction results with actual clinical intervention operations, resulting in the model remaining only in the retrospective research stage. Summary of the Invention

[0003] In response to the problems in the related art, the present invention proposes a tumor recurrence risk prediction method and system based on electronic medical record data to overcome the above-mentioned technical problems existing in the existing related art.

[0004] To solve the above technical problems, the present invention is achieved through the following technical solutions: The present invention provides a method for predicting tumor recurrence risk based on electronic medical record data, comprising the following steps: S1. Collect and pre-process the historical electronic medical records of several patients corresponding to multiple age groups; S2. Classify the pre-processed data in S1 and the corresponding patient's tumor recurrence data by age group and tumor type and set the corresponding recurrence risk level to obtain a historical recurrence risk level classification data set and a historical processed electronic medical record classification data set; S3. Constructing a tumor recurrence risk level mapping model using the historical recurrence risk level classification dataset and the historical tumor recurrence classification dataset; S4. Pre-processing the electronic medical record data of the patient to be predicted and inputting it into the corresponding tumor recurrence risk level mapping model for mapping to obtain current recurrence risk level data; S5. If the current recurrence risk level data is 0, perform tumor marker parameter measurement. If the recurrence risk level data corresponding to the measurement result is always 0, no intervention is required; otherwise, select the corresponding intervention process; if the current recurrence risk level data is not 0, directly select the corresponding intervention process; S6. Perform treatment intervention on the current patient to be predicted according to the intervention process selected in the two cases in S5 until the recurrence risk level data reaches 0.

[0005] Preferably, the S1 comprises the following steps: S11. Setting several age groups of tumor treatment populations to obtain a treatment population age group set; based on the treatment population age group set, collecting historical electronic medical record data and tumor types of several patients corresponding to each age group to obtain a historical electronic medical record data set and a historical tumor type data set; S12, pre-processing the historical electronic medical record dataset to obtain a historical processed electronic medical record dataset; By integrating heterogeneous electronic medical record data by age group, the analytical bias problem caused by the confounding of age characteristics in traditional research is solved, enabling the risk prediction model to more accurately capture the evolution patterns of tumors at different stages of the life cycle.

[0006] Preferably, the S12 includes the following steps: S121. Classify the historical electronic medical record data set to obtain a classified historical electronic medical record data set; the data classification may be based on structured, semi-structured, and unstructured data; S122, preprocessing the data of each category in the historical electronic medical record classification data set; after the preprocessing is completed, obtaining the historical processed electronic medical record data set; The term mapping and unit unification of structured and semi-structured data eliminates data heterogeneity among medical institutions and makes electronic medical records from different sources directly comparable; by extracting clinical entities and associations, free text is converted into analyzable semantic networks, supporting the construction of subsequent tumor recurrence risk prediction models.

[0007] Preferably, said S2 comprises the following steps: S21. Collecting data on whether a tumor recurred and the time of recurrence for each patient corresponding to the historical electronic medical record dataset after obtaining their respective historical electronic medical record data, in conjunction with the historical tumor type dataset, to obtain a historical tumor recurrence dataset. S22. Classifying the tumor recurrence data and electronic medical record data corresponding to different types of tumors and patients of different age groups in the historical tumor recurrence dataset and the historical processed electronic medical record dataset based on the historical tumor type dataset and the age group set of the treated population, to obtain a historical tumor recurrence classification dataset and a historical processed electronic medical record classification dataset. S23. According to the historical tumor recurrence classification dataset, setting a corresponding recurrence risk level for each patient in S11 to obtain a historical recurrence risk level classification dataset; The dataset classified by age group and tumor type provides targeted data for the subsequent construction of specific age groups and tumor types, which improves the accuracy compared to the general model.

[0008] Preferably, the S23 includes the following steps: S231, setting multiple recurrence time intervals and corresponding risk levels corresponding to each type of tumor, to obtain a tumor recurrence time interval set and a tumor recurrence risk level set; S232. If the recurrence flag data in the historical tumor recurrence classification dataset is negative, the corresponding recurrence risk is set to 0; otherwise, based on the tumor recurrence time interval set and the tumor recurrence risk level set, the recurrence time data in the historical tumor recurrence classification dataset is classified and the corresponding tumor recurrence risk level data is obtained to obtain a historical recurrence risk level classification dataset; Based on the mapping relationship between time intervals and risk levels, discrete recurrence data are converted into quantifiable and comparable risk levels, supporting clinical rapid identification of high-risk patients who require priority intervention; the classification processing of non-recurring patients and recurring patients avoids data noise interference and improves model training efficiency.

[0009] Preferably, the step S3 includes the following steps: S31. Constructing mapping models between tumor recurrence risk level data and electronic medical record data corresponding to different types of tumors and patients of different age groups, using the historical recurrence risk level classification dataset and the historical processed electronic medical record classification dataset, to obtain a tumor recurrence risk level mapping model set; Differentiated modeling based on tumor type and age stratification effectively solves the problem of insufficient prediction accuracy of traditional single models for tumors with different biological characteristics such as breast cancer and prostate cancer, and improves the accuracy of identifying high-risk patients; secondly, through the dynamic mapping of electronic medical record data and recurrence risk level, it can integrate multiple key indicators such as pathological grading, molecular typing, and treatment response to provide clinicians with real-time risk warnings, which improves efficiency compared to traditional manual evaluation.

[0010] Preferably, the mapping model in the tumor recurrence risk level mapping model set in S31 adopts a graph neural network model; Through node and edge relationship modeling, it can simultaneously process structured data, unstructured text and image features in electronic medical records, and realize automatic cross-modal feature extraction and interaction; it supports the use of attributes such as patient age and tumor type as node features, and treatment records and complications as edge relationships to construct dynamic heterogeneous graphs; through graph convolutional layers or graph attention networks, it captures the temporal dependencies of patients' historical medical records and the topological structural characteristics of tumor evolution; through graph embedding technology, discrete electronic medical records are encoded into continuous vectors, retaining the causal logic between clinical events; and using message passing mechanisms to simulate the propagation of recurrence risks in similar patient groups, improving the predictive robustness of small sample high-risk medical records.

[0011] Preferably, the S4 comprises the following steps: S41, combining the tumor recurrence risk level set, the treatment population age group set, and the historical tumor type data set, respectively setting treatment intervention processes for patients of different age groups with different types of tumors at different recurrence risk levels to obtain a tumor treatment intervention process set; S42: Obtain electronic medical record data of the current patient to be predicted, recorded as current electronic medical record data; pre-process the current electronic medical record data using the method of S12 to obtain current processed electronic medical record data; input the current processed electronic medical record data into a tumor recurrence risk level mapping model set corresponding to the age group and tumor type of the current patient to be predicted, and perform mapping to obtain current recurrence risk level data; Through graph neural network modeling of multi-dimensional data, dynamic quantitative assessment of recurrence risk can be achieved, overcoming the static limitations of the traditional TNM staging system; a three-dimensional intervention prediction and matching mechanism of risk level-tumor type-age is established, and the tumor recurrence risk level is mapped to facilitate subsequent matching of corresponding intervention processes.

[0012] Preferably, the S5 comprises the following steps: S51. Set a mapping relationship between the current prediction result observation period, the current tumor marker parameter and parameter value interval, and the tumor recurrence risk level data to obtain a current parameter interval level mapping relationship set; if the current recurrence risk level data is 0, proceed to S52; otherwise, proceed to S53; S52: Measure various parameters of the tumor markers of the patient to be predicted based on the current prediction result observation period and the current tumor marker parameter type; and if the tumor recurrence risk level data corresponding to the tumor marker parameters of the patient to be predicted is always 0, the prediction is terminated without intervention; Otherwise, the maximum tumor recurrence risk level data corresponding to the tumor marker parameters of the current patient to be predicted and the currently processed electronic medical record data are added to the historical recurrence risk level classification data set and the historical processed electronic medical record classification data set, respectively, and S31 is repeated; Then, based on the maximum tumor recurrence risk level data corresponding to the tumor marker parameters of the current patient to be predicted, the age group of the current patient to be predicted, and the tumor type, a corresponding tumor treatment intervention process is selected from the tumor treatment intervention process set to obtain a first current tumor treatment intervention process, and then proceed to S61; S53, selecting a corresponding tumor treatment intervention process from the tumor treatment intervention process set based on the current recurrence risk level data, the age group of the current patient to be predicted, and the tumor type, to obtain a second current tumor treatment intervention process, and proceeding to S62; By establishing a quantitative mapping relationship between tumor marker parameter intervals and recurrence risk levels, a transition from static single detection to dynamic continuous assessment is achieved; when the marker parameters exceed the threshold, historical data backtracking is automatically triggered, and risk stratification is optimized by combining multi-dimensional data such as age and tumor type (re-optimizing data, retraining and retesting the model), avoiding the lag of the traditional TNM staging system; differentiated treatment pathways are automatically matched according to risk levels to ensure that low-risk patients avoid overtreatment and high-risk patients receive timely and precise intervention.

[0013] Preferably, the S6 comprises the following steps: S61: Setting a first current intervention period; performing a tumor treatment intervention on the current patient to be predicted within the first current intervention period using a first current tumor treatment intervention process; after the treatment is completed, collecting tumor marker parameters of the current patient to be predicted and obtaining corresponding tumor recurrence risk level data, thereby obtaining first current post-intervention recurrence risk level data; When the first current post-intervention recurrence risk level data is greater than 0, obtaining a corresponding tumor treatment intervention process according to the first current post-intervention recurrence risk level data and the tumor treatment intervention process set and repeating the treatment intervention process in S61 until the first current post-intervention recurrence risk level data is equal to 0; S62: Setting a second current intervention period; performing a tumor treatment intervention on the current patient to be predicted within the second current intervention period using the second current tumor treatment intervention process; after the treatment is completed, collecting tumor marker parameters of the current patient to be predicted and obtaining corresponding tumor recurrence risk level data to obtain second post-current intervention recurrence risk level data; When the second current post-intervention recurrence risk level data is greater than 0, obtaining a corresponding tumor treatment intervention process according to the second current post-intervention recurrence risk level data and the tumor treatment intervention process set and repeating the treatment intervention process in S62 until the second current post-intervention recurrence risk level data is equal to 0; Automatically match the intervention process according to the initial risk level to avoid overtreatment of low-risk patients, while ensuring that high-risk patients receive timely intensive intervention; through real-time feedback of recurrence risk level data, dynamic adjustment of treatment plans is achieved until biological indicators are completely relieved; when marker parameters exceed the safety threshold, process iteration is immediately triggered, and tumor marker parameters are used as the core evaluation indicator of intervention effect, replacing simple imaging assessment, to capture microscopic lesion changes earlier.

[0014] A tumor recurrence risk prediction system based on electronic medical record data includes a historical electronic medical record data collection and processing module, a historical patient tumor data classification module, a historical recurrence risk level setting module, a recurrence risk level mapping model construction module, a current patient recurrence risk level mapping module, a current patient treatment intervention process determination module, and a current patient treatment intervention module.

[0015] The present invention has the following beneficial effects: 1. This invention structures historical electronic medical record data and establishes a recurrence risk mapping model, enabling personalized recurrence risk assessment based on multidimensional features and significantly reducing subjective judgment errors. The risk level stratification mechanism automatically distinguishes patients requiring emergency intervention from those requiring routine monitoring, avoiding over-medicalization or delayed treatment, and is particularly suitable for chronic diseases such as tumors that require long-term management. Through periodic marker measurement and risk level feedback, a closed-loop management system of "assessment-intervention-reassessment" is formed, meeting the needs of continuous optimization of clinical diagnosis and treatment.

[0016] 2. This invention addresses the analytical bias problem caused by the confounding of age characteristics in traditional research by integrating heterogeneous electronic medical record data by age group, enabling the risk prediction model to more accurately capture the evolution of tumors at different life cycle stages. The preprocessing process can eliminate data heterogeneity between medical institutions and enhance the generalization capability of subsequent modeling. In addition, preprocessing allows for direct subsequent use of preprocessed multi-dimensional features, avoiding repeated data cleaning work.

[0017] 3. The differentiated modeling based on tumor type and age stratification in this invention effectively solves the problem of insufficient prediction accuracy of traditional single models for tumors with different biological characteristics such as breast cancer and prostate cancer, and improves the accuracy of identifying high-risk patients; secondly, through the dynamic mapping of electronic medical record data and recurrence risk level, it can integrate multiple key indicators such as pathological grading, molecular typing, and treatment response to provide clinicians with real-time risk warnings, which is more efficient than traditional manual assessment; finally, the integrated learning architecture of the model set can adaptively optimize the prediction logic, while maintaining a high sensitivity while controlling the false positive rate below a reasonable range, significantly reducing the risk of over-medicalization.

[0018] Of course, any product implementing the present invention does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0020] Figure 1 This is a schematic diagram of the overall process of a method for predicting tumor recurrence risk based on electronic medical record data according to the present invention; Figure 2 A schematic diagram of the process of constructing a tumor recurrence risk level mapping model according to the present invention; Figure 3 This is a flow chart of the present invention for performing tumor treatment intervention on a patient to be predicted with an initial recurrence risk of 0; Figure 4 This is a flow chart of the present invention for performing tumor treatment intervention on a patient to be predicted and having an initial recurrence risk that is not zero; Figure 5 This is a module schematic diagram of a tumor recurrence risk prediction system based on electronic medical record data of the present invention. DETAILED DESCRIPTION

[0021] The following will clearly and completely describe the technical solutions in the embodiments of the invention in conjunction with the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of the invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0022] Example 1: Please refer to Figures 1-4 This embodiment provides a method for predicting tumor recurrence risk based on electronic medical record data, comprising the following steps: S1. Collect and pre-process the historical electronic medical records of several patients corresponding to multiple age groups; Said S1 comprises the following steps: S11. Setting several age groups of tumor treatment populations to obtain a treatment population age group set; based on the treatment population age group set, collecting historical electronic medical record data and tumor types of several patients corresponding to each age group to obtain a historical electronic medical record data set and a historical tumor type data set; S12, pre-processing the historical electronic medical record dataset to obtain a historical processed electronic medical record dataset; The S12 includes the following steps: S121. Classify the historical electronic medical record data set to obtain a classified historical electronic medical record data set; the data classification may be based on structured, semi-structured, and unstructured data; S122, preprocessing the data of each category in the historical electronic medical record classification data set; after the preprocessing is completed, obtaining the historical processed electronic medical record data set; Exemplarily, the specific process of the pre-processing is as follows: For structured and semi-structured data: Match fixed template fields (such as "pathological diagnosis: {conclusion}") through regular expressions, extract diagnostic conclusions and grading information, convert JSON / XML format inspection reports (such as endoscopy descriptions) into two-dimensional table structures, map descriptive terms to standard codes (such as "CA" is unified to "cancer", "Tumor" is mapped to ICD-O code), convert non-standard units (such as "5 thousand / mm³" → "5×10 9 / L”) and link multiple examination records of the same patient to build a time chain (such as comparison of postoperative pathology and review reports); This can be addressed by developing specialized parsers (e.g., breast cancer pathology template ≠ lung cancer template); For unstructured data (free text, images, audio, etc.): (1) Text data (medical records, chief complaints, etc.) Using models like BERT, we extract clinical entities (e.g., the symptom "persistent chest pain" and the drug "aspirin") and establish entity associations (e.g., "diabetes" → "insulin therapy"). We also remove meaningless symbols (e.g., garbled text) and merge synonyms (e.g., "myocardial infarction" = "myocardial infarction"). (2) Image data (CT / MRI / DICOM files) Convert non-DICOM images (such as JPG ultrasound images) to standard DICOM format, use the CLAHE algorithm to enhance the visibility of low-contrast lesions, and use 3D-CNN to extract radiomic features (such as tumor texture and edge sharpness); (3) Time series signal data (ECG, EEG) Perform filtering and denoising (wavelet transform to remove myoelectric interference) and segment heartbeat / brainwave bands; Term mapping between structured and semi-structured data (e.g., “CA” to “cancer”) and unit unification (e.g., “5 thousand / mm³” to “5×10 9 / L"), eliminating data heterogeneity between medical institutions and making electronic medical records from different sources directly comparable. By extracting clinical entities (such as drugs and symptoms) and their relationships, free text is converted into analyzable semantic networks, supporting the construction of subsequent tumor recurrence risk prediction models. Image data standardization (DICOM format conversion) and feature extraction (3D-CNN texture analysis) provide high-quality input for the construction of subsequent tumor recurrence risk prediction models. For example, quantification of tumor edge features can improve the accuracy of malignancy discrimination. Time series data filtering (such as ECG denoising) and segmentation processing can accurately capture changes in pathological features and assist doctors in identifying early abnormal signals (such as arrhythmias). The construction of a time series chain of multiple patient examination records supports dynamic evaluation of treatment response (such as postoperative follow-up comparison), providing a basis for personalized adjustment plans. By integrating heterogeneous electronic medical record data by age group, the analytical bias caused by the confounding of age characteristics in traditional research is resolved, enabling risk prediction models to more accurately capture the evolution of tumors at different lifecycle stages. Preprocessing processes (such as missing value filling and feature standardization) can eliminate data heterogeneity between medical institutions and improve the generalization ability of subsequent modeling. In addition, preprocessing allows for direct access to preprocessed multi-dimensional features (such as pathology text, imaging features, and test indicators) in subsequent data cleaning, avoiding repeated data cleaning. S2. Classify the pre-processed data in S1 and the corresponding patient's tumor recurrence data by age group and tumor type and set the corresponding recurrence risk level to obtain a historical recurrence risk level classification data set and a historical processed electronic medical record classification data set; The S2 comprises the following steps: S21. Collecting information on whether a tumor recurred and the time of recurrence for each patient corresponding to the historical electronic medical record dataset after obtaining their respective historical electronic medical record data, in conjunction with the historical tumor type dataset, to obtain a historical tumor recurrence dataset; Exemplarily, the historical tumor recurrence dataset is as follows: S22. Classify the tumor recurrence data and electronic medical record data corresponding to different types of tumors and patients of different age groups in the historical tumor recurrence dataset and the historical processed electronic medical record dataset based on the historical tumor type dataset and the treated population age group set, to obtain a historical tumor recurrence classification dataset and a historical processed electronic medical record classification dataset; S23. According to the historical tumor recurrence classification dataset, setting a corresponding recurrence risk level for each patient in S11 to obtain a historical recurrence risk level classification dataset; The S23 includes the following steps: S231, setting multiple recurrence time intervals and corresponding risk levels corresponding to each type of tumor, to obtain a tumor recurrence time interval set and a tumor recurrence risk level set; S232. If the recurrence flag data in the historical tumor recurrence classification dataset is negative, the corresponding recurrence risk is set to 0; otherwise, based on the tumor recurrence time interval set and the tumor recurrence risk level set, the recurrence time data in the historical tumor recurrence classification dataset is classified and the corresponding tumor recurrence risk level data is obtained to obtain a historical recurrence risk level classification dataset; Based on the mapping relationship between time intervals (e.g., the high-risk period is six months after surgery) and risk levels (e.g., low / medium / high), discrete recurrence data is converted into quantifiable and comparable risk levels, supporting the rapid identification of high-risk patients who require priority intervention in clinics; the classification of non-recurrent patients (risk set to 0) and recurrent patients avoids data noise interference and improves model training efficiency; different time interval thresholds are set for different tumor types (e.g., Luminal A type and triple-negative type of breast cancer), and follow-up frequencies can be tailored (e.g., high-risk type requires a review every three months); the correlation analysis between risk level and treatment response provides a basis for dynamic adjustment of treatment plans. By linking treatment records and recurrence time data in electronic medical records, we quantify differences in recurrence risk across different tumor types and age groups, providing data support for the subsequent construction of a recurrence risk prediction model based on real-world tumor recurrence data. Furthermore, datasets categorized by age group and tumor type provide targeted data for the subsequent construction of models tailored to specific age groups and tumor types (such as elderly breast cancer recurrence prediction models), improving accuracy by 20% to 30% compared to general models. By establishing a dataset categorized by historical recurrence risk levels, we provide labeled data for the subsequent construction of a tumor recurrence risk prediction model. S3. Constructing a tumor recurrence risk level mapping model using the historical recurrence risk level classification dataset and the historical tumor recurrence classification dataset; The S3 includes the following steps: S31. Constructing mapping models between tumor recurrence risk level data and electronic medical record data corresponding to different types of tumors and patients of different age groups, using the historical recurrence risk level classification dataset and the historical processed electronic medical record classification dataset, to obtain a tumor recurrence risk level mapping model set; The S31 includes the following steps: S311. For each group of classified data in the historical recurrence risk level classification dataset and the historical processed electronic medical record classification dataset, a corresponding training data ratio is set to obtain a training data ratio set; and according to the training data ratio set, each group of classified data in the historical recurrence risk level classification dataset and the historical processed electronic medical record classification dataset is divided to obtain a historical recurrence risk level classification training dataset, a historical processed electronic medical record classification training dataset, a historical recurrence risk level classification test dataset, and a historical processed electronic medical record classification test dataset. Construct an initial graph neural network model between the tumor recurrence risk level data corresponding to different types of tumors and patients of different age groups and electronic medical record data to obtain an initial graph neural network model set; The structure of the initial graph neural network model is as follows: 1. Graph Embedding Layer: This layer uses a Graph Attention Network (GAT) or a Graph Convolutional Network (GCN) as a foundational module to process nodes (e.g., clinical indicators, genetic features) and edges (e.g., treatment temporal relationships) in the patient-disease association graph. The input dimension is dynamically adjusted based on the number of electronic medical record features (e.g., structured data such as age and tumor stage are mapped to node attributes). 2. Spatiotemporal feature extraction layer: This includes the temporal convolutional layer (TCN), which captures the dynamic evolution of a patient's medical history. The convolution kernel width is set to 3-5 to cover the critical time window. The bidirectional LSTM layer processes unevenly spaced medical event sequences. The number of hidden units is typically set to 64-128. 3. Multimodal fusion layer: This layer integrates image features (tumor region features extracted by CNN) with graph embedding output through a cross-attention mechanism, with the weight matrix dimension matching the feature dimensions of each modality. 4. Parameter configuration: as shown in the following table: 5. Activation function selection ‌Graph Convolutional Layer‌: ‌GELU‌ (Gaussian Error Linear Unit) is preferred, as it is more adaptable to the sparsity of medical data than ReLU; ‌Fully Connected Layer‌: ‌Swish‌ activation function is used before the output layer to balance gradient vanishing and computational efficiency; S312. Setting a training error threshold set; based on the historical recurrence risk level classification training data set and the historical processed electronic medical record classification training data set, input each set of historical processed electronic medical record classification training data as training data and each corresponding set of historical recurrence risk level classification training data as training labels into the corresponding initial graph neural network model for training; in conjunction with the training error threshold set, when the training error of each initial graph neural network model is less than the corresponding training error threshold during the training process, the training is stopped to obtain a trained graph neural network model set; S313. Set a test accuracy threshold set; based on the historical recurrence risk level classification test data set and the historical processed electronic medical record classification test data set, input each set of historical processed electronic medical record classification test data as test data and each corresponding set of historical recurrence risk level classification test data as test labels into the corresponding trained graph neural network model in S312 for testing; after the test is completed, obtain the test accuracy data set; In conjunction with the test accuracy threshold set, when there is accuracy data in the test accuracy data set that is less than the corresponding test accuracy threshold, return to S312 to continue training the trained graph neural network model corresponding to the accuracy data until there is no accuracy data in the test accuracy data set that is less than the corresponding test accuracy threshold; Through the relationship modeling of nodes and edges, it can simultaneously process structured data (test indicators), unstructured text (disease course description) and imaging features in electronic medical records, and realize automatic cross-modal feature extraction and interaction; it supports the use of attributes such as patient age and tumor type as node features, and treatment records and complications as edge relationships to construct dynamic heterogeneous graphs; through graph convolutional layers (GCN) or graph attention networks (GAT), it captures the temporal dependencies of patients' historical medical records and the topological structural characteristics of tumor evolution; through graph embedding technology, discrete electronic medical records are encoded into continuous vectors, retaining the causal logic between clinical events; using the message passing mechanism to simulate the propagation of recurrence risk in similar patient groups, improving the predictive robustness of small sample high-risk medical records Differentiated modeling based on tumor type and age stratification effectively solves the problem of insufficient prediction accuracy of traditional single models for tumors with different biological characteristics such as breast cancer and prostate cancer, and improves the identification accuracy of high-risk patients by more than 40%; secondly, through the dynamic mapping of electronic medical record data and recurrence risk level, it can integrate multiple key indicators such as pathological grading, molecular typing, and treatment response to provide clinicians with real-time risk warnings, which is 6 times more efficient than traditional manual evaluation; finally, the integrated learning architecture of the model set can adaptively optimize the prediction logic, while maintaining 85% sensitivity while controlling the false positive rate below 12%, significantly reducing the risk of over-medicalization; it is suitable for guiding the formulation of individualized follow-up plans for tumor types with delayed recurrence characteristics such as HR+ / HER2- breast cancer. S4. Pre-processing the electronic medical record data of the patient to be predicted and inputting it into the corresponding tumor recurrence risk level mapping model for mapping to obtain current recurrence risk level data; The S4 comprises the following steps: S41, combining the tumor recurrence risk level set, the treatment population age group set, and the historical tumor type data set, respectively setting treatment intervention processes for patients of different age groups with different types of tumors at different recurrence risk levels to obtain a tumor treatment intervention process set; S42: Obtain electronic medical record data of the current patient to be predicted, recorded as current electronic medical record data; pre-process the current electronic medical record data using the method of S12 to obtain current processed electronic medical record data; input the current processed electronic medical record data into a tumor recurrence risk level mapping model set corresponding to the age group and tumor type of the current patient to be predicted, and perform mapping to obtain current recurrence risk level data; By modeling graph neural networks using multi-dimensional data (tumor type, age, and medical history), dynamic quantitative assessment of recurrence risk is achieved, overcoming the static limitations of the traditional TNM staging system. A three-dimensional intervention prediction and matching mechanism based on risk level, tumor type, and age is established. By mapping the tumor recurrence risk level, it facilitates subsequent matching of corresponding intervention processes. This supports dynamic adjustment of treatment plans based on changes in risk level during the treatment process, forming a continuous improvement cycle of "assessment-intervention-reassessment." S5. If the current recurrence risk level data is 0, perform tumor marker parameter measurement. If the recurrence risk level data corresponding to the measurement result is always 0, no intervention is required; otherwise, select the corresponding intervention process; if the current recurrence risk level data is not 0, directly select the corresponding intervention process; The S5 comprises the following steps: S51. Set a mapping relationship between the current prediction result observation period, the current tumor marker parameter and parameter value interval, and the tumor recurrence risk level data to obtain a current parameter interval level mapping relationship set; if the current recurrence risk level data is 0, proceed to S52; otherwise, proceed to S53; S52: Measure various parameters of the tumor markers of the patient to be predicted based on the current prediction result observation period and the current tumor marker parameter type; and if the tumor recurrence risk level data corresponding to the tumor marker parameters of the patient to be predicted is always 0, the prediction is terminated without intervention; Otherwise, the maximum tumor recurrence risk level data corresponding to the tumor marker parameters of the current patient to be predicted and the currently processed electronic medical record data are added to the historical recurrence risk level classification data set and the historical processed electronic medical record classification data set, respectively, and S31 is repeated; Then, based on the maximum tumor recurrence risk level data corresponding to the tumor marker parameters of the current patient to be predicted, the age group of the current patient to be predicted, and the tumor type, a corresponding tumor treatment intervention process is selected from the tumor treatment intervention process set to obtain a first current tumor treatment intervention process, and then proceed to S61; S53, selecting a corresponding tumor treatment intervention process from the tumor treatment intervention process set based on the current recurrence risk level data, the age group of the current patient to be predicted, and the tumor type, to obtain a second current tumor treatment intervention process, and proceeding to S62; For example, take a female patient with lung adenocarcinoma as an example: Initial assessment: A 58-year-old female patient with lung adenocarcinoma underwent postoperative follow-up. Serum CEA was 2.8 ng / mL (<4.26 ng / mL), CYFRA21-1 was 5.2 ng / mL (<9.03 ng / mL), and the risk level was mapped to 03. Initiate a 3-month observation period and continuously monitor marker levels; At the third follow-up, CYFRA21-1 rose to 11.5 ng / mL (>9.03 ng / mL), triggering a high-risk level (level 2); The system automatically matches the platinum-containing doublet chemotherapy regimen (age 50-65 years, lung adenocarcinoma, risk level 2); By establishing a quantitative mapping relationship between tumor marker parameter intervals and recurrence risk levels (such as CEA>4.26 ng / mL, CYFRA21-1>9.03 ng / mL and other thresholds are associated with high-risk levels), achieving a shift from static single testing to dynamic continuous assessment; when a marker parameter exceeds a threshold, historical data backtesting is automatically triggered, and risk stratification is optimized by combining multi-dimensional data such as age and tumor type, avoiding the lag of the traditional TNM staging system; differentiated treatment pathways (direct intervention and continuous monitoring) are automatically matched according to risk level (0 / non-0), ensuring that low-risk patients avoid overtreatment and high-risk patients receive timely precision intervention; in addition, through iterative updates of historical data sets, the training and test data of the tumor recurrence risk prediction model are continuously enriched, thereby greatly improving the predictive accuracy and clinical usability of the tumor recurrence risk level mapping model after training and testing; using tumor markers, such as obtaining the ROC curve characteristics of markers such as CEA and CA125 (AUC>0.7), combined with cross-validation of electronic medical record data, significantly reduces the false negative / positive rate; secondly, through a risk level pre-screening mechanism, medical resources are preferentially allocated to high-risk groups, shortening the decision-making cycle from testing to intervention; S6. Performing treatment intervention on the current patient to be predicted according to the intervention process selected in the two cases in S5 until the recurrence risk level data reaches 0; The S6 comprises the following steps: S61: Setting a first current intervention period; performing a tumor treatment intervention on the current patient to be predicted within the first current intervention period using a first current tumor treatment intervention process; after the treatment is completed, collecting tumor marker parameters of the current patient to be predicted and obtaining corresponding tumor recurrence risk level data, thereby obtaining first current post-intervention recurrence risk level data; When the first current post-intervention recurrence risk level data is greater than 0, obtaining a corresponding tumor treatment intervention process according to the first current post-intervention recurrence risk level data and the tumor treatment intervention process set and repeating the treatment intervention process in S61 until the first current post-intervention recurrence risk level data is equal to 0; S62: Setting a second current intervention period; performing a tumor treatment intervention on the current patient to be predicted within the second current intervention period using the second current tumor treatment intervention process; after the treatment is completed, collecting tumor marker parameters of the current patient to be predicted and obtaining corresponding tumor recurrence risk level data to obtain second post-current intervention recurrence risk level data; When the second current post-intervention recurrence risk level data is greater than 0, obtaining a corresponding tumor treatment intervention process according to the second current post-intervention recurrence risk level data and the tumor treatment intervention process set and repeating the treatment intervention process in S62 until the second current post-intervention recurrence risk level data is equal to 0; For example, take patient A as an example: Baseline assessment: Patient A (68 years old, colorectal cancer, CA19-9=380U / mL, risk level 3): Intervention duration: 12 weeks; Regimen: FOLFOXIRI (oxaliplatin + irinotecan + 5-FU); Dynamic adjustment: CA19-9 = 210 U / mL at week 8 (risk level 2): ​​additional bevacizumab was added, and the final CA19-9 = 35 U / mL (risk level 0); Automatically match the intervention process according to the initial risk level to avoid overtreatment of low-risk patients, while ensuring that high-risk patients receive timely intensive intervention; through real-time feedback of recurrence risk level data (repeat the intervention process when >0), dynamic adjustment of the treatment plan is achieved until the biological indicators are completely relieved (risk level = 0); preset intervention cycles (first / second intervention periods) are combined with dynamic monitoring of markers to avoid the lag of traditional fixed treatment models; when the marker parameters exceed the safety threshold, the process iteration is immediately triggered, and tumor marker parameters are used as the core evaluation indicator of the intervention effect, replacing simple imaging assessment to capture microscopic lesion changes earlier.

[0023] Example 2: Please refer to Figure 5 This embodiment discloses a tumor recurrence risk prediction system based on electronic medical record data. The system can implement the method of the above embodiment, including a historical electronic medical record data collection and processing module, a historical patient tumor data classification module, a historical recurrence risk level setting module, a recurrence risk level mapping model construction module, a current patient recurrence risk level mapping module, a current patient treatment intervention process determination module, and a current patient treatment intervention module; The historical electronic medical record data collection and processing module collects and pre-processes the historical electronic medical record data of a number of patients corresponding to multiple age groups to obtain a historical processed electronic medical record data set; The historical patient tumor data classification module classifies the historical processed electronic medical record dataset and the tumor recurrence data of the corresponding patients according to age group and tumor type to obtain a historical tumor recurrence classification dataset and a historical processed electronic medical record classification dataset; The historical recurrence risk level setting module sets the recurrence risk level of each patient in S1 corresponding to the historical tumor recurrence classification data set to obtain a historical recurrence risk level classification data set; The recurrence risk level mapping model construction module uses the historical recurrence risk level classification data set and the historical tumor recurrence classification data set to construct a tumor recurrence risk level mapping model; The current patient recurrence risk level mapping module pre-processes the electronic medical record data of the current patient to be predicted and inputs it into the corresponding tumor recurrence risk level mapping model for mapping to obtain current recurrence risk level data; The current patient treatment intervention process determination module performs periodic tumor marker parameter measurements if the current recurrence risk level data is 0. If the recurrence risk level data corresponding to the measurement results is always 0, no intervention is required; otherwise, intervention is performed and a corresponding intervention process is selected to obtain a first current tumor treatment intervention process; if the current recurrence risk level data is not 0, intervention is performed and a corresponding intervention process is selected to obtain a second current tumor treatment intervention process; The current patient treatment intervention module performs treatment intervention on the current patient to be predicted according to the first current tumor treatment intervention process and the second current tumor treatment intervention process until the recurrence risk level data is 0.

[0024] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0025] The preferred embodiments of the invention disclosed above are intended only to help illustrate the invention. These preferred embodiments do not exhaust all details, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention.

Claims

1. A method for predicting tumor recurrence risk based on electronic medical record data, characterized in that: The following steps are involved: S1. Collect and pre-process the historical electronic medical records of several patients corresponding to multiple age groups; S2. Classify the pre-processed data in S1 and the corresponding patient's tumor recurrence data by age group and tumor type and set the corresponding recurrence risk level to obtain a historical recurrence risk level classification data set and a historical processed electronic medical record classification data set; S3. Constructing a tumor recurrence risk level mapping model using the historical recurrence risk level classification dataset and the historical tumor recurrence classification dataset; S4. Pre-processing the electronic medical record data of the patient to be predicted and inputting it into the corresponding tumor recurrence risk level mapping model for mapping to obtain current recurrence risk level data; S5. If the current recurrence risk level data is 0, perform tumor marker parameter measurement. If the recurrence risk level data corresponding to the measurement result is always 0, no intervention is required; otherwise, select the corresponding intervention process; If the current recurrence risk level data is not 0, directly select the corresponding intervention process; S6. Perform treatment intervention on the current patient to be predicted according to the intervention process selected in the two cases in S5 until the recurrence risk level data reaches 0.

2. The method for predicting tumor recurrence risk based on electronic medical record data according to claim 1, characterized in that: Said S1 comprises the following steps: S11. Setting several age groups of tumor treatment populations to obtain a treatment population age group set; based on the treatment population age group set, collecting historical electronic medical record data and tumor types of several patients corresponding to each age group to obtain a historical electronic medical record data set and a historical tumor type data set; S12, pre-processing the historical electronic medical record dataset to obtain a historical processed electronic medical record dataset; The S12 includes the following steps: S121. Classify the historical electronic medical record data set to obtain a classified historical electronic medical record data set; the data classification may be based on structured, semi-structured, and unstructured data; S122 , preprocessing the data of each category in the historical electronic medical record classification data set; after the preprocessing is completed, obtaining a historical processed electronic medical record data set.

3. The method for predicting tumor recurrence risk based on electronic medical record data according to claim 2, characterized in that: The S2 comprises the following steps: S21. Collecting data on whether a tumor recurred and the time of recurrence for each patient corresponding to the historical electronic medical record dataset after obtaining their respective historical electronic medical record data, in conjunction with the historical tumor type dataset, to obtain a historical tumor recurrence dataset. S22. Classifying the tumor recurrence data and electronic medical record data corresponding to different types of tumors and patients of different age groups in the historical tumor recurrence dataset and the historical processed electronic medical record dataset based on the historical tumor type dataset and the age group set of the treated population, to obtain a historical tumor recurrence classification dataset and a historical processed electronic medical record classification dataset. S23. According to the historical tumor recurrence classification dataset, a corresponding recurrence risk level is set for each patient in S11 to obtain a historical recurrence risk level classification dataset.

4. The method for predicting tumor recurrence risk based on electronic medical record data according to claim 3, characterized in that: The S23 includes the following steps: S231, setting multiple recurrence time intervals and corresponding risk levels corresponding to each type of tumor, to obtain a tumor recurrence time interval set and a tumor recurrence risk level set; S232. If the recurrence flag data in the historical tumor recurrence classification data set is negative, the corresponding recurrence risk is set to 0; otherwise, in combination with the tumor recurrence time interval set and the tumor recurrence risk level set, the recurrence time data in the historical tumor recurrence classification data set is classified and the corresponding tumor recurrence risk level data is obtained to obtain a historical recurrence risk level classification data set.

5. The method for predicting tumor recurrence risk based on electronic medical record data according to claim 4, characterized in that: The S3 includes the following steps: S31. In conjunction with the historical recurrence risk level classification dataset and the historical processed electronic medical record classification dataset, mapping models between tumor recurrence risk level data corresponding to different types of tumors and patients of different age groups and electronic medical record data are constructed to obtain a tumor recurrence risk level mapping model set.

6. The method for predicting tumor recurrence risk based on electronic medical record data according to claim 5, characterized in that: The mapping model in the tumor recurrence risk level mapping model set described in S31 adopts a graph neural network model.

7. The method for predicting tumor recurrence risk based on electronic medical record data according to claim 4, characterized in that: The S4 comprises the following steps: S41, combining the tumor recurrence risk level set, the treatment population age group set, and the historical tumor type data set, respectively setting treatment intervention processes for patients of different age groups with different types of tumors at different recurrence risk levels to obtain a tumor treatment intervention process set; S42. Obtain the electronic medical record data of the current patient to be predicted, recorded as the current electronic medical record data; pre-process the current electronic medical record data using the method of S12 to obtain the current processed electronic medical record data; input the current processed electronic medical record data into the tumor recurrence risk level mapping model set corresponding to the age group and tumor type of the current patient to be predicted, and map it to obtain the current recurrence risk level data.

8. The method for predicting tumor recurrence risk based on electronic medical record data according to claim 7, characterized in that: The S5 comprises the following steps: S51. Set a mapping relationship between the current prediction result observation period, the current tumor marker parameter and parameter value interval, and the tumor recurrence risk level data to obtain a current parameter interval level mapping relationship set; if the current recurrence risk level data is 0, proceed to S52; otherwise, proceed to S53; S52: Measure various parameters of the tumor markers of the patient to be predicted based on the current prediction result observation period and the current tumor marker parameter type; and if the tumor recurrence risk level data corresponding to the tumor marker parameters of the patient to be predicted is always 0, the prediction is terminated without intervention; Otherwise, the maximum tumor recurrence risk level data corresponding to the tumor marker parameters of the current patient to be predicted and the currently processed electronic medical record data are added to the historical recurrence risk level classification data set and the historical processed electronic medical record classification data set, respectively, and S31 is repeated; Then, based on the maximum tumor recurrence risk level data corresponding to the tumor marker parameters of the current patient to be predicted, the age group of the current patient to be predicted, and the tumor type, a corresponding tumor treatment intervention process is selected from the tumor treatment intervention process set to obtain a first current tumor treatment intervention process, and then proceed to S61; S53. Select a corresponding tumor treatment intervention process from the tumor treatment intervention process set based on the current recurrence risk level data, the age group of the current patient to be predicted, and the tumor type, to obtain a second current tumor treatment intervention process, and proceed to S62.

9. The method for predicting tumor recurrence risk based on electronic medical record data according to claim 8, characterized in that: The S6 comprises the following steps: S61: Setting a first current intervention period; performing a tumor treatment intervention on the current patient to be predicted within the first current intervention period using a first current tumor treatment intervention process; after the treatment is completed, collecting tumor marker parameters of the current patient to be predicted and obtaining corresponding tumor recurrence risk level data, thereby obtaining first current post-intervention recurrence risk level data; When the first current post-intervention recurrence risk level data is greater than 0, obtaining a corresponding tumor treatment intervention process according to the first current post-intervention recurrence risk level data and the tumor treatment intervention process set and repeating the treatment intervention process in S61 until the first current post-intervention recurrence risk level data is equal to 0; S62: Setting a second current intervention period; performing a tumor treatment intervention on the current patient to be predicted within the second current intervention period using the second current tumor treatment intervention process; after the treatment is completed, collecting tumor marker parameters of the current patient to be predicted and obtaining corresponding tumor recurrence risk level data to obtain second post-current intervention recurrence risk level data; When the second current post-intervention recurrence risk level data is greater than 0, the corresponding tumor treatment intervention process is obtained according to the second current post-intervention recurrence risk level data and the tumor treatment intervention process set and the above treatment intervention process in S62 is repeated until the second current post-intervention recurrence risk level data is equal to 0.

10. A tumor recurrence risk prediction system based on electronic medical record data, characterized in that: Used to implement a tumor recurrence risk prediction method based on electronic medical record data as described in any one of claims 1-9.

Citation Information

Patent Citations

  • Extraction process of total gingerols

    CN109675002A

  • NSCLC patient postoperative short-period recurrence metastasis risk evaluating system

    CN110111892A

  • Tumor recurrence risk assessment and early warning system based on data analysis

    CN119361146A

  • Intelligent classification prediction method, system and equipment for tumor treatment and protection

    CN119650072A

  • Chest tumor prognosis prediction method and system based on data analysis

    CN120048514A

Cited By

  • Muscle weakness identification method and system for critical patient

    CN120727262A

  • A method and system for identifying muscle weakness in critically ill patients

    CN120727262B

  • Risk prediction and evaluation method and device for pulmonary arterial hypertension

    CN121421572A