Mammary nodule malignant transformation risk prediction system and method, terminal, medium and product

Abstract

The invention provides a mammary nodule malignant transformation risk prediction system and method, a terminal, a medium and a product. The system comprises a multi-source data acquisition module used for acquiring multi-source mammary gland related medical data of a target mammary nodule patient and preprocessing the multi-source mammary gland related medical data; the key prediction feature screening module is used for carrying out key feature screening on the preprocessed data to obtain mammary gland key prediction feature data; and the prediction module is used for inputting the breast key feature data into a pre-constructed breast nodule risk prediction model to obtain a corresponding breast nodule malignant transformation risk prediction result. According to the method, the related medical data of the multi-source mammary gland is integrated, and key feature screening and the mammary nodule risk prediction model are combined, so that the limitation of traditional single-dimensional evaluation is broken through, and early warning of malignant transformation of the mammary nodule is realized.

Description

Technical Field

[0001] This application relates to the field of model prediction technology, and in particular to a system, method, terminal, medium and product for predicting the risk of malignant transformation of breast nodules. Background Technology

[0002] Breast nodules are common breast lesions in women, and some benign nodules carry a risk of malignant transformation. Early and accurate identification of these high-risk nodules is crucial for reducing the incidence and mortality of breast cancer. Existing methods for assessing breast nodule risk mainly rely on single-dimensional indicators: for example, the ultrasound BI-RADS classification is based solely on imaging characteristics, which suffers from strong subjective dependence and a high rate of missed detection of early, small lesions; while pathological biopsy is the gold standard, it is an invasive procedure and cannot dynamically assess the transformation risk of nodules; traditional clinical scoring systems (such as the Gail model) only incorporate limited demographic and medical history characteristics, resulting in insufficient accuracy in early warning (AUC is mostly between 0.7 and 0.8).

[0003] With the development of precision medicine, multi-omics data fusion and machine learning technologies have provided new directions for risk warning, but existing technologies still have shortcomings: First, the data dimension is single and does not fully integrate multi-source information such as imaging, pathology, and molecular markers; second, the model generalization ability is weak, and it is mostly based on single-center small sample data and lacks external validation; third, the warning is delayed and it is difficult to predict the trend of malignant transformation in advance. Summary of the Invention

[0004] In view of the shortcomings of the prior art described above, the purpose of this application is to provide a system, method, terminal, medium and product for predicting the risk of malignant transformation of breast nodules, so as to solve the problem that the existing technology has a single data dimension and is difficult to predict the trend of malignant transformation of breast nodules in advance.

[0005] To achieve the above and other related objectives, the first aspect of this application provides a breast nodule malignant transformation risk prediction system, comprising: a multi-source data acquisition module for acquiring and preprocessing multi-source breast-related medical data of a target breast nodule patient; a key prediction feature screening module for screening key features of the preprocessed data to obtain breast key prediction feature data; and a prediction module for inputting the breast key feature data into a pre-constructed breast nodule risk prediction model to obtain the corresponding breast nodule malignant transformation risk prediction result.

[0006] In some embodiments of the first aspect of this application, the multi-source breast-related medical data includes at least: clinical baseline data, imaging feature data, and molecular marker data.

[0007] In some embodiments of the first aspect of this application, the preprocessing includes: missing value processing, outlier removal based on interquartile range, and data standardization; wherein, the missing value processing includes: mode imputation of categorical variables in clinical baseline data, multiple imputation of continuous variables in clinical baseline data, and K-nearest neighbor imputation of imaging feature data and molecular marker data.

[0008] In some embodiments of the first aspect of this application, key feature screening is performed on preprocessed data to obtain key breast cancer predictive feature data, including: inputting the preprocessed data into a LASSO regression model for dimensionality reduction to obtain dimensionality-reduced feature data; wherein, the LASSO regression model introduces an L1 regularization term; using mutual information, calculating the correlation between the dimensionality-reduced feature data and a pre-set target variable, and screening out key breast cancer predictive feature data from the dimensionality-reduced feature data according to a pre-set correlation threshold; wherein, the target variable is the dimensionality-reduced feature data of patients with malignant transformation outcomes.

[0009] In some embodiments of the first aspect of this application, the training process of the breast nodule risk prediction model specifically includes: obtaining key breast prediction feature data of multiple patients and labeling whether malignant transformation has occurred to construct an initial dataset; performing preprocessing and key feature screening on the initial dataset in sequence to obtain a key breast prediction feature dataset; using the training dataset, training a random forest model and an XGBoost model respectively, and fusing the trained random forest model and the trained XGBoost model to obtain a breast nodule risk prediction model.

[0010] In some embodiments of the first aspect of this application, a breast nodule risk prediction model is obtained by fusing a trained random forest model and an XGBoost model, including: determining the AUC values ​​of the trained random forest model and the XGBoost model, and assigning weights to the trained random forest model and the XGBoost model respectively based on the AUC values ​​of the two models; and performing weighted fusion of the outputs of the two models based on the weights of the trained random forest model and the trained XGBoost model to obtain the breast nodule risk prediction model; wherein the output after weighted fusion is the output of the breast nodule risk prediction model.

[0011] To achieve the above and other related objectives, a second aspect of this application provides a method for predicting the risk of malignant transformation of breast nodules, comprising: acquiring multi-source breast-related medical data of a patient with a target breast nodule and preprocessing it; screening key features of the preprocessed data to obtain key predictive feature data of the breast; and inputting the key predictive feature data of the breast into a pre-constructed breast nodule risk prediction model to obtain the corresponding prediction result of the risk of malignant transformation of the breast nodule.

[0012] To achieve the above and other related objectives, a third aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for predicting the risk of malignant transformation of breast nodules.

[0013] To achieve the above and other related objectives, a fourth aspect of this application provides a computer program product comprising computer program code, which, when executed on a computer, enables the computer to implement the method for predicting the risk of malignant transformation of breast nodules.

[0014] To achieve the above and other related objectives, a fifth aspect of this application provides an electronic terminal, including a memory, a processor, and a computer program stored in the memory; the processor executes the computer program to implement the method for predicting the risk of malignant transformation of breast nodules.

[0015] As described above, the breast nodule malignant transformation risk prediction system, method, terminal, medium and product of this application have the following beneficial effects: By integrating multi-source breast-related medical data and combining key feature screening and breast nodule risk prediction model, this application not only breaks through the limitations of traditional single-dimensional assessment, but also realizes early warning of breast nodule malignant transformation. Attached Figure Description

[0016] Figure 1 The diagram shown is a schematic block diagram of a breast nodule malignant transformation risk prediction system according to an embodiment of this application.

[0017] Figure 2 The diagram shown is a flowchart illustrating a method for predicting the risk of malignant transformation of breast nodules in one embodiment of this application.

[0018] Figure 3 The diagram shown is a structural schematic of an electronic terminal according to an embodiment of this application. Detailed Implementation

[0019] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0020] In the embodiments of this application, the terms "first" and "second" are used to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that the terms "first" and "second" do not necessarily imply that they are different.

[0021] It should be noted that, in the embodiments of this application, the words "exemplary" or "for example" indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner.

[0022] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0023] To facilitate understanding of the embodiments of this application, firstly, in conjunction with Figure 1 Detailed explanation. Figure 1 A schematic block diagram of a breast nodule malignant transformation risk prediction system according to an embodiment of the present invention is shown. The breast nodule malignant transformation risk prediction system 100 in this embodiment includes:

[0024] The multi-source data acquisition module 101 is used to acquire multi-source breast-related medical data of the target breast nodule patient and to preprocess the multi-source breast-related medical data of the target breast nodule patient;

[0025] The key prediction feature filtering module 102 is used to filter key features from the preprocessed data to obtain key prediction feature data of breast cancer.

[0026] The prediction module 103 is used to input key breast feature data into a pre-constructed breast nodule risk prediction model to obtain the corresponding breast nodule malignant transformation risk prediction results.

[0027] It should be understood that the module division in the embodiments of this application is illustrative and only represents a logical functional division. In actual implementation, there may be other division methods. Furthermore, the functional modules in the various embodiments of this application can be integrated into a single processor, exist as separate physical entities, or be integrated into a single module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0028] In one embodiment, the breast nodule malignancy transformation risk prediction system is based on a B / S architecture design and supports multi-terminal access. It should be understood that B / S architecture is an abbreviation for Browser / Server, a distributed software architecture model based on the Internet. Its core feature is that the client only needs to install a general-purpose browser, without the need to develop a dedicated client program; all business logic, data processing, and resource storage are centralized on the server side.

[0029] In one embodiment, the multi-source data acquisition module has a multi-source data import interface that supports uploading files in formats such as DICOM (radiography), HL7 (electronic medical records), and CSV (test indicators). The multi-source data acquisition module also has built-in data validity verification rules that automatically identify and remove invalid data with incorrect formats or missing fields, while recording the data collection time and source.

[0030] To facilitate understanding, this invention first explains the training process of the breast nodule risk prediction model:

[0031] Multi-source breast-related medical data were acquired from multiple patients with breast nodules, and each data point was labeled to indicate whether it had a malignant transformation outcome, in order to construct an initial dataset. The multi-source breast-related medical data included at least clinical baseline data, imaging feature data, and molecular marker data. Optionally, pathological data were also included.

[0032] Specifically, clinical baseline data is extracted from desensitized electronic medical records, including but not limited to patient age, family history of breast cancer, age of menarche, menopausal status, number of births, breastfeeding history, history of hormone replacement therapy, time of first discovery of nodules, and trend of size changes;

[0033] The imaging feature data consists of quantitative and qualitative features extracted from imaging data such as ultrasound (nodule size, borders, shape, internal echoes, calcification type, blood flow signal, BI-RADS classification), mammography (percentage of dense breast tissue, nodule density, calcification distribution), and MRI (ADC value, enhancement pattern, time-signal intensity curve type). Quantitative features (e.g., maximum nodule diameter, ADC value) are directly extracted numerically, while qualitative features (e.g., border clarity, calcification type) are converted into numerical data using an encoding method.

[0034] Pathological data are extracted from the examination data of puncture biopsy, including but not limited to pathological type (e.g., fibroma, hyperplastic nodule, dysplasia, etc.), degree of cellular atypia (mild / moderate / severe), Ki-67 index (proliferative activity index), etc.

[0035] Molecular marker data were obtained through laboratory testing, including but not limited to the expression levels of serum tumor markers (CEA, CA15-3) and microRNAs (miR-21, miR-155). The HER2 gene amplification status was encoded by immunohistochemical results (0 / 1+ / 2+ / 3+).

[0036] Further, the initial dataset is preprocessed. Preprocessing includes missing value handling, outlier removal, and data standardization. Specifically, missing value handling includes: mode imputation for categorical variables (e.g., family history of breast cancer) in clinical baseline data, and multiple imputation for continuous variables (e.g., age) in clinical baseline data. Missing values ​​in imaging feature data and molecular marker data are imputed using K-nearest neighbor (KNN). It should be understood that mode imputation means replacing all missing values ​​of a variable with the most frequent category (i.e., the mode) in the dataset. Multiple imputation (MICE, Multivariate Imputation by Chained Equations) is the mainstream statistical method for handling missing data. Its core is to iteratively generate multiple complete datasets through chain equations, analyze them separately, and then merge the results to preserve data variability and avoid underestimation of variance. K-nearest neighbor (KNN) imputation is used for missing values ​​in imaging feature and molecular marker data. It should be understood that K-nearest neighbor imputation means finding the K most similar samples to the sample containing missing values, and then using the weighted average or mode of the corresponding features of these K samples to fill the missing values. For example, K can be set to 5.

[0037] For continuous characteristic variables (such as nodule size and biomarker concentration), outliers are removed by calculating the interquartile range (ICM). The ICM is calculated using the following formula:

[0038] (1)

[0039] Where IQR stands for interquartile range. The upper quartile refers to the value at the 75th percentile (75% of the data is less than or equal to the upper quartile) after sorting the data of the same type of variable in the initial dataset from smallest to largest. ), The lower quartile refers to the value at the 25th percentile (25% of the data is less than or equal to) after sorting the data of the same type in the initial dataset from smallest to largest. It should be understood that different types of variables have different interquartile ranges, and outlier removal is performed for each type of variable based on its corresponding interquartile range.

[0040] Next, the variable values ​​in the dataset are less than or greater than Data removal.

[0041] Then, the dataset after missing value processing and outlier removal is standardized to obtain an intermediate training dataset. Standardization is performed using the Z-score standardization formula (Formula 2) to eliminate the impact of dimensional differences on model training:

[0042] (2)

[0043] in, These are the variables in the dataset after missing value processing and outlier removal. For standardized variables, The mean, The standard deviation is denoted as .

[0044] Furthermore, the intermediate training dataset is sequentially subjected to LASSO regression dimensionality reduction and mutual information filtering to form a dataset of key breast cancer prediction features. Specifically, the intermediate training dataset is input into the LASSO regression model for dimensionality reduction, resulting in a dimensionality-reduced dataset. An L1 regularization term is added to the LASSO regression model to compress feature coefficients, thereby achieving feature selection and dimensionality reduction. It should be understood that LASSO (Least Absolute Shrinkage and Selection Operator) is a linear regression improvement technique, the core of which lies in adding an L1 regularization term to the model's loss function. For example, the search range of the penalty coefficient λ of the L1 regularization term can be set to [0.001, 0.1], and 10-fold cross-validation is used to determine the optimal λ value. In this case, features with non-zero regression coefficients are retained, while redundant features with coefficients of 0 are removed.

[0045] Multi-source breast-related medical data of patients with breast nodules exhibiting malignant transformation outcomes were acquired and subjected to the aforementioned preprocessing and LASSO regression dimensionality reduction to obtain dimensionality-reduced feature data of these patients, which was then used as the target variable. The correlation between each feature data point in the dimensionality-reduced dataset and the target variable was calculated, using a correlation calculation formula based on existing mutual information methods. The correlation value ranged from [0, 1], with a higher value indicating a stronger correlation. Next, based on a pre-set correlation threshold, key breast prediction feature data were selected from each dimensionality-reduced feature data point to obtain the final key breast prediction feature dataset. For example, the correlation threshold could be set to 0.1 to retain features with a mutual information value ≥ 0.1.

[0046] Furthermore, based on the key breast cancer prediction feature dataset, multiple training subsets are generated through bootstrap sampling. It should be understood that bootstrap sampling is a resampling statistical method that generates a large number of pseudo-samples (bootstrap samples) by repeatedly sampling the original dataset with replacement, thereby simulating the sampling distribution of the data without relying on traditional theoretical assumptions. The number of training subsets is the same as the number of decision trees in the random forest. Each training subset trains one decision tree, and the final prediction results are ensembled using a voting method to obtain the trained random forest model. For example, the number of decision trees can be set to 150, the maximum tree depth to 8, and the minimum number of samples for node splits to 5.

[0047] The XGBoost model was trained using a dataset of key breast cancer prediction features to obtain a trained XGBoost model. It should be understood that the XGBoost model is a gradient boosting decision tree algorithm. For example, during model training, the learning rate can be set to 0.08, the number of trees to 120, the maximum tree depth to 5, and the regularization parameter to 0.1.

[0048] Then, weights are assigned based on the AUC values ​​of the two trained models on the validation set, and the two models are fused to obtain the breast nodule risk prediction model. Specifically, in use, the outputs of the two models are weighted and fused to obtain the final output of the breast nodule risk prediction model. It should be understood that the method for obtaining the validation set can refer to the method for obtaining the breast key prediction feature dataset mentioned above, and will not be repeated here. The AUC (Area Under the ROC Curve) value is the gold standard for evaluating the performance of binary classification models. Essentially, it measures the model's ability to distinguish between positive and negative samples, and its value ranges from [0,1]. For example, if Random Forest AUC = 0.88 and XGBoost AUC = 0.90, then the weights are 0.47 and 0.53, respectively.

[0049] In one specific embodiment, five-fold cross-validation is employed. The aforementioned key breast cancer prediction feature dataset is randomly divided into five parts, with four parts used as the training set and one part used for five iterations. The average validation index is calculated. Data from breast nodule patients at another independent medical institution (sample size ≥ 200 cases, follow-up time ≥ 3 years) is selected and processed as described above as an external validation set to evaluate the model's generalization ability on new data. If the AUC, sensitivity, specificity, and accuracy of both cross-validation and external validation meet the preset thresholds, the model is determined as the final breast nodule risk prediction model; otherwise, the model returns to the feature selection step to readjust the thresholds or adjusts the model hyperparameters for retraining.

[0050] The following explains the working process of the breast nodule malignant transformation risk prediction system:

[0051] Acquire multi-source breast-related medical data from patients with target breast nodules. This data should include at least clinical baseline data, imaging features, and molecular markers to ensure comprehensive early warning. Additionally, if the patient underwent a biopsy, the multi-source breast-related medical data should also include pathological data.

[0052] Specifically, clinical baseline data is extracted from desensitized electronic medical records, including but not limited to patient age, family history of breast cancer, age of menarche, menopausal status, number of births, breastfeeding history, history of hormone replacement therapy, time of first discovery of nodules, and trend of size changes;

[0053] The imaging feature data consists of quantitative and qualitative features extracted from imaging data such as ultrasound (nodule size, borders, shape, internal echoes, calcification type, blood flow signal, BI-RADS classification), mammography (percentage of dense breast tissue, nodule density, calcification distribution), and MRI (ADC value, enhancement pattern, time-signal intensity curve type). Quantitative features (e.g., maximum nodule diameter, ADC value) are directly extracted numerically, while qualitative features (e.g., border clarity, calcification type) are converted into numerical data using an encoding method.

[0054] Pathological data are extracted from the examination data of puncture biopsy, including but not limited to pathological type (e.g., fibroma, hyperplastic nodule, dysplasia, etc.), degree of cellular atypia (mild / moderate / severe), Ki-67 index (proliferative activity index), etc.

[0055] Molecular marker data were obtained through laboratory testing, including but not limited to the expression levels of serum tumor markers (CEA, CA15-3) and microRNAs (miR-21, miR-155). The HER2 gene amplification status was encoded by immunohistochemical results (0 / 1+ / 2+ / 3+).

[0056] Further, preprocessing includes missing value handling, outlier removal, and data standardization. Specifically, categorical variables (e.g., family history of breast cancer) in the clinical baseline data are imputed using the mode, while continuous variables (e.g., age) are imputed using multiple imputation. Missing values ​​in imaging feature data and molecular marker data are imputed using K-nearest neighbor (KNN). It should be noted that in the prediction phase, mode imputation means replacing all missing values ​​of the variable with the most frequent category (i.e., the mode) in the initial dataset. K-nearest neighbor (KNN) imputation can also refer to the initial dataset described above. Outlier removal can use the threshold calculated above. Data standardization also follows the method described in the above embodiment and will not be repeated here.

[0057] Furthermore, the preprocessed data is subjected to LASSO regression dimensionality reduction and mutual information screening to obtain key predictive features of breast cancer. The LASSO regression dimensionality reduction and mutual information screening can be performed as described in the above embodiments, and will not be repeated here.

[0058] Finally, the key predictive features of breast tissue are input into the breast nodule risk prediction model to obtain the corresponding prediction results of the risk of malignant transformation of breast nodules.

[0059] In one specific embodiment, the prediction result of the risk of malignant transformation of breast nodules can be represented by a malignant transformation risk value (0~100 points), and the risk value is divided into levels (low risk: 0~30 points, intermediate risk: 31~60 points, high risk: 61~100 points). The risk score radar chart and key risk feature ranking are displayed through visualization units, and a structured report containing risk assessment conclusions and clinical intervention recommendations (such as regular follow-up examinations, further biopsy, surgical intervention) is generated, which supports export in PDF format.

[0060] In one embodiment, the breast nodule malignant transformation risk prediction system periodically (e.g., every 6 months) receives new patients' multi-source breast-related medical data and malignant transformation outcomes, and uses an incremental learning algorithm to update the model parameters without retraining the entire model, thereby improving update efficiency and ensuring model timeliness.

[0061] This invention integrates multi-source data from clinical baselines, imaging features, pathological biopsies, and molecular markers, combined with precise feature selection using LASSO regression and mutual information methods, to construct an ensemble learning model that weights random forest and XGBoost. Validated using five-fold cross-validation and external independent datasets (core performance metrics reach AUC≥0.85, sensitivity≥85%), this model overcomes the limitations of traditional single-dimensional assessments and solves the problems of weak generalization and delayed early warning in existing models. Clinically, it can identify high-risk breast nodules for malignant transformation 3-6 months in advance, through risk stratification (low / The system provides individualized intervention recommendations for patients at medium / high risk, avoiding overtreatment of low-risk patients while securing an intervention window for high-risk patients. This reduces the risk of missed diagnosis and progression of breast cancer, while also decreasing reliance on invasive procedures and improving patient adherence to follow-up. The system supports multi-terminal access and automated processing, and is equipped with encrypted data storage to ensure privacy and security. It is easy to operate and promote in medical institutions, optimizing clinical treatment pathways and medical resource allocation. Furthermore, it helps build a closed-loop system for early breast cancer diagnosis encompassing screening, assessment, intervention, and follow-up, demonstrating significant technological innovation, clinical practicality, and public health value.

[0062] Figure 2 This is a flowchart illustrating the method for predicting the risk of malignant transformation of breast nodules provided in this application. Figure 2 As shown, methods for predicting the risk of malignant transformation of breast nodules include:

[0063] Step S201: Obtain multi-source breast-related medical data of the target breast nodule patient and preprocess them.

[0064] Step S202: Perform key feature screening on the preprocessed data to obtain key predictive feature data for breast cancer.

[0065] Step S203: Input the key breast feature data into the pre-constructed breast nodule risk prediction model to obtain the corresponding breast nodule malignant transformation risk prediction results.

[0066] It should be understood that the execution process of each step has been described in detail in the above system embodiments, and will not be repeated here for the sake of brevity.

[0067] In one embodiment, the multi-source breast-related medical data includes at least: clinical baseline data, imaging feature data, and molecular marker data.

[0068] In one embodiment, the preprocessing includes: missing value processing, outlier removal based on interquartile range, and data standardization; wherein, the missing value processing includes: mode imputation of categorical variables in clinical baseline data, multiple imputation of continuous variables in clinical baseline data, and K-nearest neighbor imputation of imaging feature data and molecular marker data.

[0069] In one embodiment, key feature screening is performed on preprocessed data to obtain key predictive feature data for breast cancer, including: inputting the preprocessed data into a LASSO regression model for dimensionality reduction to obtain dimensionality-reduced feature data; wherein, the LASSO regression model introduces an L1 regularization term; using mutual information, calculating the correlation between the dimensionality-reduced feature data and a pre-set target variable, and screening out key predictive feature data for breast cancer from the dimensionality-reduced feature data according to a pre-set correlation threshold; wherein, the target variable is the dimensionality-reduced feature data of patients with malignant transformation outcomes.

[0070] In one embodiment, the training process of the breast nodule risk prediction model specifically includes: obtaining key breast prediction feature data of multiple patients and labeling whether malignant transformation has occurred to construct an initial dataset; performing preprocessing and key feature screening on the initial dataset in sequence to obtain a key breast prediction feature dataset; using the training dataset, training a random forest model and an XGBoost model respectively, and fusing the trained random forest model and the trained XGBoost model to obtain a breast nodule risk prediction model.

[0071] In one embodiment, a breast nodule risk prediction model is obtained by fusing a trained random forest model and an XGBoost model, including: determining the AUC values ​​of the trained random forest model and the XGBoost model, and assigning weights to the trained random forest model and the XGBoost model respectively based on the AUC values ​​of the two models; and performing weighted fusion of the outputs of the two models based on the weights of the trained random forest model and the trained XGBoost model to obtain the breast nodule risk prediction model; wherein the output after weighted fusion is the output of the breast nodule risk prediction model.

[0072] Figure 3 This is a schematic block diagram of the electronic terminal provided in an embodiment of this application. Figure 3As shown, the electronic terminal includes at least one processor 301, a memory 302, at least one network interface 303, and a user interface 305. The various components in the device are coupled together via a bus system 304. It is understood that the bus system 304 is used to implement communication between these components. In addition to a data bus, the bus system 304 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in… Figure 3 The general will label all buses as bus systems.

[0073] The user interface 305 may include a monitor, keyboard, mouse, trackball, clicker, button, touchpad, or touch screen.

[0074] It is understood that memory 302 can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM) or programmable read-only memory (PROM), used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM) and synchronous static random access memory (SSRAM). The memories described in the embodiments of this invention are intended to include, but are not limited to, these and any other suitable categories of memory.

[0075] In this embodiment of the invention, the memory 302 is used to store various types of data to support the operation of the electronic terminal 300. Examples of this data include: any executable program for operation on the electronic terminal 300, such as the operating system 3021 and application programs 3022; the operating system 3021 contains various system programs, such as the framework layer, core library layer, driver layer, etc., for implementing various basic services and handling hardware-based tasks. The application program 3022 may contain various applications, such as media players, browsers, etc., for implementing various application services. The method for predicting the risk of malignant transformation of breast nodules provided in this embodiment of the invention can be included in the application program 3022.

[0076] The methods disclosed in the above embodiments of the present invention can be applied to processor 301, or implemented by processor 301. Processor 301 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the integrated logic circuit of the hardware in processor 301 or by instructions in the form of software. The processor 301 may be a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 301 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. General-purpose processor 301 may be a microprocessor or any conventional processor, etc. The steps of the accessory optimization method provided in the embodiments of the present invention can be directly reflected as being executed by a hardware decoding processor, or being executed by a combination of hardware and software modules in the decoding processor. The software module may be located in a storage medium, which is located in memory. The processor reads the information in the memory and combines it with its hardware to complete the steps of the aforementioned method.

[0077] In an exemplary embodiment, the electronic terminal 300 may be used by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), or complex programmable logic devices (CPLDs) to execute the aforementioned method.

[0078] According to the method provided in the embodiments of this application, this application also provides a computer program product, which includes: computer program code, which, when run on a computer, causes the computer to execute... Figure 2 The method for predicting the risk of malignant transformation of breast nodules in the illustrated embodiment.

[0079] According to the method provided in the embodiments of this application, this application also provides a computer-readable storage medium storing program code, which, when executed on a computer, causes the computer to perform... Figure 2 The method for predicting the risk of malignant transformation of breast nodules in the illustrated embodiment.

[0080] As used in this specification, the terms "component," "module," "system," etc., are used to refer to computer-related entities, hardware, firmware, combinations of hardware and software, software, or software in execution. For example, a component can be, but is not limited to, a process running on a processor, a processor, an object, an executable file, an execution thread, a program, and / or a computer. As illustrated, applications running on computing devices and computing devices can both be components. One or more components may reside in a process and / or an execution thread, and components may be located on a single computer and / or distributed among two or more computers. Furthermore, these components can be executed from various computer-readable media on which various data structures are stored. Components can communicate, for example, via local and / or remote processes based on signals having one or more data packets (e.g., data from two components interacting with another component between a local system, a distributed system, and / or a network, such as the Internet interacting with other systems via signals).

[0081] Those skilled in the art will recognize that the various illustrative logical blocks and steps described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this application.

[0082] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0083] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0084] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0085] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0086] In the above embodiments, the functions of each functional unit can be implemented entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. A computer program product includes one or more computer instructions (programs). When the computer program instructions (programs) are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., high-density digital video discs, DVDs), or semiconductor media (e.g., solid-state disks, SSDs, etc.).

[0087] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0088] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0089] In summary, this application provides a system, method, terminal, medium, and product for predicting the risk of malignant transformation of breast nodules. The system includes: a multi-source data acquisition module for acquiring and preprocessing multi-source breast-related medical data of patients with target breast nodules; a key predictive feature screening module for screening key features of the preprocessed data to obtain key predictive feature data of the breast; and a prediction module for inputting the key feature data of the breast into a pre-constructed breast nodule risk prediction model to obtain the corresponding prediction result of the risk of malignant transformation of breast nodules. This application, by integrating multi-source breast-related medical data and combining key feature screening with a breast nodule risk prediction model, overcomes the limitations of traditional single-dimensional assessment and achieves early warning of malignant transformation of breast nodules. Therefore, this application effectively overcomes the various shortcomings of existing technologies and has high industrial application value.

[0090] The above embodiments are merely illustrative of the principles and effects of this application and are not intended to limit this application. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of this application. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in this application should still be covered by the claims of this application.

Claims

1. A system for predicting the risk of malignant transformation of breast nodules, characterized in that, include: The multi-source data acquisition module is used to acquire and preprocess multi-source breast-related medical data of patients with target breast nodules. The key prediction feature filtering module is used to filter key features from the preprocessed data to obtain key prediction feature data for breast cancer. The prediction module is used to input key breast feature data into a pre-built breast nodule risk prediction model to obtain the corresponding breast nodule malignancy transformation risk prediction results.

2. The breast nodule malignant transformation risk prediction system according to claim 1, characterized in that, The multi-source breast-related medical data includes at least: clinical baseline data, imaging feature data, and molecular marker data.

3. The breast nodule malignant transformation risk prediction system according to claim 2, characterized in that, The preprocessing includes: missing value processing, outlier removal based on interquartile range, and data standardization; wherein, the missing value processing includes: mode imputation for categorical variables in clinical baseline data, multiple imputation for continuous variables in clinical baseline data, and K-nearest neighbor imputation for imaging feature data and molecular marker data.

4. The breast nodule malignant transformation risk prediction system according to claim 2, characterized in that, The preprocessed data was subjected to key feature screening to obtain key breast cancer prediction feature data, including: The preprocessed data is input into the LASSO regression model for dimensionality reduction to obtain the dimensionality-reduced feature data; wherein, the LASSO regression model introduces an L1 regularization term; Using mutual information, the correlation between the dimensionality-reduced feature data and the pre-defined target variable is calculated, and key predictive feature data of breast cancer is selected from the dimensionality-reduced feature data according to the pre-defined correlation threshold; wherein, the target variable is the dimensionality-reduced feature data of patients with malignant transformation outcomes.

5. The breast nodule malignant transformation risk prediction system according to claim 1, characterized in that, The training process of the breast nodule risk prediction model specifically includes: We obtained key predictive features of breast cancer from multiple patients and labeled them to indicate whether malignant transformation had occurred, thus constructing an initial dataset. The initial dataset is preprocessed and key features are filtered sequentially to obtain a dataset of key breast cancer prediction features. Using the training dataset, we trained a random forest model and an XGBoost model respectively, and then fused the trained random forest model and the trained XGBoost model to obtain a breast nodule risk prediction model.

6. The breast nodule malignant transformation risk prediction system according to claim 5, characterized in that, The trained random forest model and the XGBoost model are fused to obtain a breast nodule risk prediction model, including: Determine the AUC values ​​of the trained random forest model and XGBoost model, and assign weights to the trained random forest model and XGBoost model respectively based on the AUC values ​​of the two models. Based on the weights of the trained random forest model and the trained XGBoost model, the outputs of the two models are weighted and fused to obtain the breast nodule risk prediction model; the output after weighted fusion is the output of the breast nodule risk prediction model.

7. A method for predicting the risk of malignant transformation of breast nodules, characterized in that, include: Acquire multi-source breast-related medical data from patients with target breast nodules and preprocess them; Key features were selected from the preprocessed data to obtain key predictive features of breast cancer. Key breast feature data are input into a pre-constructed breast nodule risk prediction model to obtain the corresponding breast nodule malignancy transformation risk prediction results.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of claim 7.

9. A computer program product, characterized in that, The computer program product includes computer program code that, when run on a computer, causes the computer to perform the method as described in claim 7.

10. An electronic terminal, comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the method of claim 7.

Images