An intelligent data analysis system before breast tumor resection
Through the intelligent data analysis system before breast tumor resection, combined with multi-source data and machine learning technology, the problem of poor breast tumor case information analysis in traditional technology is solved, and the precise positioning and personalized treatment strategies of breast tumors are realized, which improves the Surgical success rate and treatment effectiveness.
Patent Information
- Application Number
- CN202411186753.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-08-28
AI Technical Summary
Traditional data analysis technology before breast tumor resection cannot uniformly quantify and analyze multi-source breast tumor case information, resulting in poor effect in prognostic breast tumor morphology and lesion analysis, and cannot provide accurate and efficient treatment-assisted decisions. .
A data intelligent analysis system before breast tumor resection was designed, including breast tumor case data preprocessing, pathological feature processing, target area image extraction, growth trend analysis, cutting risk prediction and treatment feature analysis modules, which are integrated through multi-source data , image segmentation, machine learning modeling and other technologies provide personalized treatment decision support.
It has achieved accurate positioning, growth trend assessment and treatment strategy formulation of breast tumors, improved the success rate of surgery and the effectiveness of treatment, and provided accurate and efficient auxiliary decision-making in the treatment of breast tumors.
Smart Images

Figure CN118711837B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of medical data processing, and in particular to an intelligent data analysis system before breast tumor resection. Background Art
[0002] Intelligent data analysis before breast tumor resection can provide a comprehensive tumor feature assessment and personalized treatment recommendations by comprehensively analyzing the patient's pathological data, imaging results and other features. It not only helps to accurately locate the location, size and growth trend of the tumor, but also predicts surgical risks, optimizes surgical plans, and provides support for patients' postoperative rehabilitation. Through intelligent data analysis, treatment strategies can be formulated more accurately to improve the success rate of medical surgery. However, traditional data analysis technology before breast tumor resection cannot combine the patient's multi-source breast tumor case information for unified quantitative analysis, resulting in poor analysis of the imaging features and pathological features of breast tumor case information for the prognosis of breast tumor morphology and pathological conditions, and cannot provide accurate and efficient prognosis of breast tumor treatment auxiliary decision-making. Summary of the invention
[0003] Based on this, the present invention provides a data intelligent analysis system before breast tumor resection to solve at least one of the above technical problems.
[0004] To achieve the above purpose, a data intelligent analysis system before breast tumor resection includes the following modules:
[0005] A breast tumor case data preprocessing module is used to obtain breast tumor case data from a medical database; preprocess the breast tumor case data to obtain standard breast tumor case data, wherein the standard breast tumor case data includes breast tumor pathology data and breast tumor imaging data;
[0006] A breast tumor pathology feature processing module is used to analyze the breast tumor pathology type based on the breast tumor pathology data and generate the breast tumor pathology type data; to calculate the breast tumor nuclear division index based on the breast tumor pathology data and generate the breast tumor nuclear division index;
[0007] A breast tumor target region image extraction module is used to perform target region image segmentation processing on breast tumor image data to generate breast tumor target region image data;
[0008] The breast tumor growth trend analysis module is used to analyze and process the breast tumor metastasis pattern based on the breast tumor target area image data to generate breast tumor metastasis pattern data; analyze the breast tumor growth trend based on the breast tumor nuclear division index and the breast tumor metastasis pattern data to generate breast tumor growth trend data;
[0009] The prognostic breast tumor cutting risk prediction module is used to construct a breast tumor cutting risk prediction model; perform prognostic breast tumor target area image analysis on the breast tumor target area image data according to the breast tumor growth trend data to generate prognostic breast tumor target area image data; transmit the prognostic breast tumor target area image data to the breast tumor cutting risk prediction model to perform prognostic breast tumor cutting risk prediction to generate prognostic breast tumor cutting risk prediction data;
[0010] The prognostic breast tumor resection information intelligent feedback module is used to perform prognostic breast tumor treatment feature analysis based on breast tumor pathological type data and prognostic breast tumor target area imaging data to obtain prognostic breast tumor treatment feature data; design prognostic breast tumor treatment auxiliary decision based on the prognostic breast tumor treatment feature data; transmit the prognostic breast tumor cutting risk prediction data and the prognostic breast tumor treatment auxiliary decision to the terminal to perform the prognostic breast tumor auxiliary resection information intelligent feedback operation.
[0011] The present invention obtains breast tumor case data from a medical database, and can collect breast tumor case data from multiple medical databases, including clinical records, imaging data and pathological data, to ensure the extensiveness and comprehensiveness of data sources. Data preprocessing is performed on the breast tumor case data, and the data preprocessing can clean and standardize data from different sources, including removing duplicate data, correcting erroneous data, and ensuring the consistency and accuracy of the data. The obtained standard breast tumor case data combines pathological data and imaging data, providing a complete and accurate basis for subsequent analysis and diagnosis. By analyzing pathological data, the system can accurately identify different types of breast tumors and the corresponding degree of deterioration, and classify and categorize them, providing an important basis for doctors' diagnosis and treatment decisions. Nuclear division index is an important indicator for evaluating the proliferation activity of tumor cells. The growth rate and potential invasiveness of the tumor are determined by this calculation, which helps to predict its development trend and the formulation of treatment strategies. Perform target region image segmentation on breast tumor image data, extract the target region of the breast tumor from the entire image, accurately locate and identify the location, shape and size of the tumor. The extracted target region image data not only helps with visual analysis in the diagnosis stage, but also serves as basic data for subsequent growth trend analysis and treatment planning. Analyzing the metastasis pattern of breast tumors based on target region image data of breast tumors, that is, the path and method of how the tumor spreads and metastasizes within the breast, is crucial for predicting the development and potential risks of tumors. In addition, combined with the nuclear division index and metastasis pattern data, it is possible to evaluate the growth rate, potential deterioration trend and treatment response of breast tumors, which helps to formulate targeted and effective treatment plans. By analyzing breast tumor growth trend data and target area imaging data, a prediction model is constructed to evaluate the risks of breast tumor resection surgery. Big data and machine learning technology are used, combined with clinical experience and scientific basis, to predict potential risks and complications during the operation. Prognostic breast tumor target area imaging analysis is performed on breast tumor target area imaging data based on breast tumor growth trend data, and the trend characteristics of breast tumors in the late stage, such as the shape, size and location of the tumor in the late stage, are further analyzed. The prognostic breast tumor target area imaging data is transmitted to the breast tumor cutting risk prediction model for prognostic breast tumor cutting risk prediction, so as to better evaluate the difficulty of the operation and the individualized risk of the patient, thereby optimizing the surgical planning and improving the safety and success rate of the operation.Prognostic breast tumor treatment feature analysis is performed based on breast tumor pathological type data and prognostic breast tumor target area imaging data, including prediction of tumor biological characteristics, individual differences and treatment responses of patients, and breast tumor cutting paths, to develop personalized treatment plans, improve treatment effectiveness and prognostic results, and thus design and provide intelligent treatment decision-making assistance plans, provide auxiliary medical advice and support before breast tumor cutting, optimize the treatment process, improve treatment accuracy and patient satisfaction, and help medical personnel adjust treatment strategies by transmitting prognostic breast tumor cutting risk prediction data and treatment assistance decisions to the terminal to ensure that patients receive the best support and management throughout the treatment process.
[0012] The beneficial effect of the present application is that the data intelligent analysis system before breast tumor resection of the present invention can accurately identify and extract the target area of breast tumor through the target area segmentation and feature extraction technology of breast tumor image data, and provide necessary data support for subsequent growth trend analysis and treatment decision-making. And the characteristics of nuclear division index and metastasis pattern of multi-source breast tumors are comprehensively analyzed, and the growth rate and diffusion tendency of tumors can be evaluated, and predictive information is provided to help adjust treatment strategies and prevent postoperative recurrence. According to breast tumor growth trend data and image analysis results, surgical risks are quantified and decision support is provided, so as to achieve data-driven treatment options. Through the prognosis breast tumor cutting risk prediction data and the prognosis breast tumor treatment auxiliary decision feedback personalized prognosis breast tumor resection information, it helps doctors optimize surgical plans and manage patients' postoperative recovery, and improves the patient's recovery status in the later stage. In order to realize the unified quantitative analysis of prognosis breast tumors in combination with multi-source breast tumor case information of patients, the image characteristics and pathological characteristics of the comprehensive breast tumor case information are better for the analysis of the morphology and pathological conditions of prognosis breast tumors, thereby providing accurate and efficient prognosis breast tumor treatment auxiliary decision-making. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 A schematic diagram of the module flow of a data intelligent analysis system before breast tumor resection according to the present invention;
[0014] Figure 2 for Figure 1 Detailed functional flow chart of the breast tumor growth trend analysis module;
[0015] Figure 3 for Figure 1 Detailed functional flow diagram of the mid-prognosis breast tumor cutting risk prediction module;
[0016] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0017] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.
[0018] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.
[0019] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.
[0020] To achieve this, please refer to Figures 1 to 3 The present invention provides a data intelligent analysis system before breast tumor resection, comprising the following modules:
[0021] A breast tumor case data preprocessing module is used to obtain breast tumor case data from a medical database; preprocess the breast tumor case data to obtain standard breast tumor case data, wherein the standard breast tumor case data includes breast tumor pathology data and breast tumor imaging data;
[0022] A breast tumor pathology feature processing module is used to analyze the breast tumor pathology type based on the breast tumor pathology data and generate the breast tumor pathology type data; to calculate the breast tumor nuclear division index based on the breast tumor pathology data and generate the breast tumor nuclear division index;
[0023] A breast tumor target region image extraction module is used to perform target region image segmentation processing on breast tumor image data to generate breast tumor target region image data;
[0024] The breast tumor growth trend analysis module is used to analyze and process the breast tumor metastasis pattern based on the breast tumor target area image data to generate breast tumor metastasis pattern data; analyze the breast tumor growth trend based on the breast tumor nuclear division index and the breast tumor metastasis pattern data to generate breast tumor growth trend data;
[0025] The prognostic breast tumor cutting risk prediction module is used to construct a breast tumor cutting risk prediction model; perform prognostic breast tumor target area image analysis on the breast tumor target area image data according to the breast tumor growth trend data to generate prognostic breast tumor target area image data; transmit the prognostic breast tumor target area image data to the breast tumor cutting risk prediction model to perform prognostic breast tumor cutting risk prediction to generate prognostic breast tumor cutting risk prediction data;
[0026] The prognostic breast tumor resection information intelligent feedback module is used to perform prognostic breast tumor treatment feature analysis based on breast tumor pathological type data and prognostic breast tumor target area imaging data to obtain prognostic breast tumor treatment feature data; design prognostic breast tumor treatment auxiliary decision based on the prognostic breast tumor treatment feature data; transmit the prognostic breast tumor cutting risk prediction data and the prognostic breast tumor treatment auxiliary decision to the terminal to perform the prognostic breast tumor auxiliary resection information intelligent feedback operation.
[0027] The present invention obtains breast tumor case data from a medical database, and can collect breast tumor case data from multiple medical databases, including clinical records, imaging data and pathological data, to ensure the extensiveness and comprehensiveness of data sources. Data preprocessing is performed on the breast tumor case data, and the data preprocessing can clean and standardize data from different sources, including removing duplicate data, correcting erroneous data, and ensuring the consistency and accuracy of the data. The obtained standard breast tumor case data combines pathological data and imaging data, providing a complete and accurate basis for subsequent analysis and diagnosis. By analyzing pathological data, the system can accurately identify different types of breast tumors and the corresponding degree of deterioration, and classify and categorize them, providing an important basis for doctors' diagnosis and treatment decisions. Nuclear division index is an important indicator for evaluating the proliferation activity of tumor cells. The growth rate and potential invasiveness of the tumor are determined by this calculation, which helps to predict its development trend and the formulation of treatment strategies. Perform target region image segmentation on breast tumor image data, extract the target region of the breast tumor from the entire image, accurately locate and identify the location, shape and size of the tumor. The extracted target region image data not only helps with visual analysis in the diagnosis stage, but also serves as basic data for subsequent growth trend analysis and treatment planning. Analyzing the metastasis pattern of breast tumors based on target region image data of breast tumors, that is, the path and method of how the tumor spreads and metastasizes within the breast, is crucial for predicting the development and potential risks of tumors. In addition, combined with the nuclear division index and metastasis pattern data, it is possible to evaluate the growth rate, potential deterioration trend and treatment response of breast tumors, which helps to formulate targeted and effective treatment plans. By analyzing breast tumor growth trend data and target area imaging data, a prediction model is constructed to evaluate the risks of breast tumor resection surgery. Big data and machine learning technology are used, combined with clinical experience and scientific basis, to predict potential risks and complications during the operation. Prognostic breast tumor target area imaging analysis is performed on breast tumor target area imaging data based on breast tumor growth trend data, and the trend characteristics of breast tumors in the late stage, such as the shape, size and location of the tumor in the late stage, are further analyzed. The prognostic breast tumor target area imaging data is transmitted to the breast tumor cutting risk prediction model for prognostic breast tumor cutting risk prediction, so as to better evaluate the difficulty of the operation and the individualized risk of the patient, thereby optimizing the surgical planning and improving the safety and success rate of the operation.Prognostic breast tumor treatment feature analysis is performed based on breast tumor pathological type data and prognostic breast tumor target area imaging data, including prediction of tumor biological characteristics, individual differences and treatment responses of patients, and breast tumor cutting paths, to develop personalized treatment plans, improve treatment effectiveness and prognostic results, and thus design and provide intelligent treatment decision-making assistance plans, provide auxiliary medical advice and support before breast tumor cutting, optimize the treatment process, improve treatment accuracy and patient satisfaction, and help medical personnel adjust treatment strategies by transmitting prognostic breast tumor cutting risk prediction data and treatment assistance decisions to the terminal to ensure that patients receive the best support and management throughout the treatment process.
[0028] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the module flow of a data intelligent analysis system before breast tumor resection of the present invention. In this embodiment, the data intelligent analysis system before breast tumor resection includes:
[0029] S1: a breast tumor case data preprocessing module, used to obtain breast tumor case data from a medical database; preprocess the breast tumor case data to obtain standard breast tumor case data, wherein the standard breast tumor case data includes breast tumor pathology data and breast tumor imaging data;
[0030] In an embodiment of the present invention, access rights to a medical database containing breast tumor case data are obtained, and structured query language (SQL) or a dedicated API interface is used to extract breast tumor case data from the medical database, including basic patient information (such as age, gender), clinical examination data (such as tumor size and location), pathological data (such as tissue type and tumor grade), imaging data (such as MRI, ultrasound and other images), etc. The extracted data is cleaned and integrated, including processing missing values, abnormal values and inconsistent data to ensure data quality and consistency, and data from different data sources are standardized into a unified format and unit. According to the characteristics of the breast tumor case data, relevant features are selected and extracted, such as key features in pathological data and imaging data, non-numeric data are encoded or converted, such as single-hot encoding or label encoding of categorical data to facilitate processing by machine learning algorithms, and numerical data are normalized or standardized so that each feature has a similar scale in terms of numerical value to obtain standard breast tumor case data, wherein the standard breast tumor case data includes processed breast tumor pathology data and breast tumor imaging data.
[0031] S2: Breast tumor pathology feature processing module, used to analyze breast tumor pathology type based on breast tumor pathology data and generate breast tumor pathology type data; calculate breast tumor nuclear division index based on breast tumor pathology data and generate breast tumor nuclear division index;
[0032] In an embodiment of the present invention, the histological features of breast tumor tissue extracted from breast tumor pathology data are preliminarily processed, including data cleaning and feature extraction. According to different features in the pathology data (such as cell morphology, tissue structure, etc.), the most representative features are selected for analysis, and the association relationship analysis of the breast tumor pathology features is performed using association analysis, and the main association features are analyzed by principal component analysis to find the similarities and differences between different cases. The case type of breast tumor is analyzed according to the association relationship and the results of principal component analysis, and breast tumor pathology type data is generated, such as classification labels or descriptive information of different types of breast tumors. According to the nuclear morphological features in the pathology data, such as nuclear size, shape, chromatin distribution, and chromatin change data, a deep learning algorithm is designed to automatically calculate the nuclear division index of breast tumor cells. According to the number and distribution characteristics of measured cell nuclei, the nuclear division index of breast tumor cells is automatically analyzed and calculated for evaluating the proliferation activity and clinical prognosis of tumors.
[0033] S3: a breast tumor target region image extraction module, used to perform target region image segmentation processing on breast tumor image data to generate breast tumor target region image data;
[0034] In an embodiment of the present invention, target region image segmentation processing is performed on breast tumor image data, such as selecting an algorithm suitable for breast tumor image segmentation, such as a segmentation algorithm based on deep learning (such as U-Net, Mask R-CNN), a region growing method or a segmentation algorithm based on image texture features, performing preliminary region marking or seed point initialization on the breast tumor image data, such as selecting an image region of a target gray value gradient interval through gray value gradient feature data of the breast tumor image data as input of the segmentation algorithm, applying the selected algorithm to accurately segment the target region of the breast tumor, ensuring that the segmentation result accurately reflects the boundary and area of the tumor, and post-processing the segmentation result, including removing unreasonable regions, filling holes, smoothing boundaries, etc., to improve the quality and continuity of the segmentation result, and generate breast tumor target region image data.
[0035] S4: Breast tumor growth trend analysis module, used to perform breast tumor metastasis pattern analysis and processing based on breast tumor target area image data to generate breast tumor metastasis pattern data; perform breast tumor growth trend analysis based on breast tumor nuclear division index and breast tumor metastasis pattern data to generate breast tumor growth trend data;
[0036] In an embodiment of the present invention, relevant biological features, such as blood vessel density, blood vessel tissue structure, etc., are extracted from the image data of the target area of the breast tumor. Based on the extracted feature data, a preset clustering algorithm (such as K-means, hierarchical clustering, etc.) is used to perform cluster analysis on the breast tumor metastasis pattern to reveal the similarities and differences between different metastasis patterns, and the breast tumor nuclear division index data is associated with the metastasis pattern data to establish an overall framework for breast tumor growth trend analysis. The breast tumor nuclear division index and the time series feature data of the breast tumor metastasis pattern data are analyzed using a time series neural network algorithm. The growth trend of the breast tumor is preliminarily analyzed and described through the time series feature data, including information such as growth rate and metastasis path. The relationship between the breast tumor metastasis pattern and the patient's survival status is further analyzed, and the potential influencing factors of the metastasis pattern on the prognosis of the breast tumor are explored. The growth trend of the breast tumor is refined and optimized and adjusted based on the potential influencing factors of the prognosis of the breast tumor, thereby obtaining the breast tumor growth trend data.
[0037] S5: A prognostic breast tumor cutting risk prediction module, which is used to construct a breast tumor cutting risk prediction model; perform prognostic breast tumor target area image analysis on breast tumor target area image data according to breast tumor growth trend data to generate prognostic breast tumor target area image data; transmit the prognostic breast tumor target area image data to the breast tumor cutting risk prediction model to perform prognostic breast tumor cutting risk prediction to generate prognostic breast tumor cutting risk prediction data;
[0038] In an embodiment of the present invention, by obtaining historical breast tumor cutting data, including clinical data related to historical breast tumor cutting, a suitable machine learning or deep learning model, such as support vector machine (SVM), random forest, convolutional neural network (CNN), etc., is selected to establish a mathematical model of the relationship between the breast tumor cutting area and the breast tumor cutting risk, which is used to predict the breast tumor cutting risk. The historical breast tumor cutting data is used for model training, and the breast tumor cutting area data of the historical breast tumor cutting data is used as the input of the model training and the cutting risk label data is used as the output of the model training, and the hyperparameters of the model are adjusted by cross-validation or grid search, etc., and finally a breast tumor cutting risk prediction model is constructed. The processed and analyzed growth trend data are obtained from the breast tumor growth trend analysis module, and these data include detailed information on the growth rate, metastasis pattern and survival relationship of the breast tumor. Next, the image data extracted by the breast tumor target area image extraction module is used to ensure that the data contains accurate breast tumor target area images. Then, the growth trend data is combined with the target area image data, and the image evolution numerical analysis of the breast tumor target area image is performed through detailed information on the growth rate, metastasis pattern and survival relationship of the breast tumor, as well as the application of image processing technology and image analysis algorithm, so as to generate prognostic breast tumor target area image data. The prognostic breast tumor target area image data is transmitted to the breast tumor cutting risk prediction model for prognostic breast tumor cutting risk prediction. The model analyzes the risk information of prognostic breast tumor cutting based on the prognostic breast tumor target area image data, such as cutting risk probability, abnormal cutting path, etc., so as to generate prognostic breast tumor cutting risk prediction data.
[0039] S6: A prognostic breast tumor resection information intelligent feedback module, which is used to perform prognostic breast tumor treatment feature analysis based on breast tumor pathological type data and prognostic breast tumor target area imaging data to obtain prognostic breast tumor treatment feature data; design prognostic breast tumor treatment auxiliary decision-making based on the prognostic breast tumor treatment feature data; transmit the prognostic breast tumor cutting risk prediction data and the prognostic breast tumor treatment auxiliary decision-making to the terminal to perform the prognostic breast tumor auxiliary resection information intelligent feedback operation.
[0040] In an embodiment of the present invention, a breast tumor pathology type characteristic analysis is performed based on the breast tumor pathology type data to obtain breast tumor pathology type characteristic data, including characteristic information such as the histological type, grade, and immunohistochemical molecular subtype of the tumor. The prognosis treatment data of the breast tumor pathology type is analyzed based on the breast tumor pathology type characteristic data, and the breast tumor is divided into different types, such as ductal carcinoma, lobular carcinoma, breast cancer, etc. Different breast tumor types are graded (such as low, medium, and high degree of differentiation), and their immunohistochemical subtypes (such as hormone receptor and HER2 expression) are determined. The prognosis treatment data of the breast tumor pathology type is analyzed based on this information and medical records. The corresponding breast tumor resection method, such as breast-conserving surgery or mastectomy, is involved in the prognosis breast tumor target area image data, so as to design the prognosis treatment data of the prognosis breast tumor target area image data. According to the breast tumor pathology type and the prognosis treatment data corresponding to the prognosis breast tumor target area image, the associated characteristics are analyzed, that is, the surgical method under the prognosis condition of the breast tumor and the corresponding later treatment data, so as to obtain the prognosis breast tumor treatment characteristic data. Design prognostic breast tumor treatment auxiliary decision-making based on prognostic breast tumor treatment characteristic data, that is, integrate the treatment plan types of all prognostic breast tumor treatment characteristic data to design prognostic breast tumor treatment auxiliary decision-making. Use the prognostic breast tumor treatment auxiliary decision-making and the surgical method corresponding to the prognostic breast tumor cutting risk prediction data as associated tags, and push the two to the terminal after associating them to perform the intelligent feedback operation of prognostic breast tumor auxiliary resection information, so that medical staff can receive the surgical method, surgical risk, prognosis and later treatment of breast tumors through the terminal.
[0041] Preferably, the breast tumor case data preprocessing module includes the following functions:
[0042] Obtain breast tumor case data from medical databases;
[0043] Performing breast tumor case information type analysis on breast tumor case data to obtain breast tumor case information type data;
[0044] Design classification case information preprocessing decisions based on breast tumor case information type data;
[0045] The breast tumor case data are preprocessed according to the classified case information preprocessing decision to obtain standard breast tumor case data.
[0046] The present invention obtains breast tumor case data from a medical database. The system ensures that the source of the data is extensive and comprehensive, integrates breast tumor case data of patients in different regions, different hospitals and different time periods, reduces bias and misleading results, and provides a reliable basis for subsequent analysis. By analyzing various types of information in breast tumor case data, such as clinical data, imaging data, pathological data, etc., the system can accurately identify and classify different types of data, classify and organize different types of data, lay the foundation for subsequent classification and analysis, and improve the availability and efficiency of data. According to the analyzed information type data, various types of preprocessing strategies and processes are formulated, and different processing methods are implemented for different types of data to ensure the quality and accuracy of the data, and appropriate preprocessing decisions are designed to effectively clean and standardize the data, eliminate noise and errors in the data, and make subsequent analysis more accurate and reliable. Data preprocessing is performed on breast tumor case data based on classified case information preprocessing decisions to standardize breast tumor case data, including data cleaning, missing value processing, data conversion, etc., to ensure data consistency and accuracy. Through optimized data preprocessing processes, the system can reduce data processing time and cost, improve data processing efficiency, and enable standard breast tumor case data to be more quickly applied to clinical and scientific research practices.
[0047] In an embodiment of the present invention, access rights to a medical database containing breast tumor case data are obtained, and structured query language (SQL) or a special API interface is used to extract breast tumor case data from the medical database, including basic patient information (such as age, gender), clinical examination data (such as tumor size and location), pathological data (such as tissue type and tumor grade), imaging data (such as MRI, ultrasound and other images), etc. According to the data source of the API interface, the breast tumor case data is subjected to a breast tumor case information type analysis to obtain breast tumor case information type data, including imaging type data and pathological type data of the breast tumor case data. According to the breast tumor case information type data, corresponding classification case information preprocessing decisions are designed, such as standardized adjustment of image grayscale and contrast of the imaging data, and image data with excessive noise is identified and cleared, and text mining is performed on the pathological type data to analyze keywords, phrases or text features, identify text anomalies and remove them. Data preprocessing is performed on breast tumor case data according to the classified case information preprocessing decision. Based on the preprocessing decisions of various types of case information in the classified case information preprocessing decision, such as the preprocessing decisions of imaging data, pathological data, etc., data preprocessing is performed uniformly to obtain standard breast tumor case data, wherein the standard breast tumor case data includes preprocessed breast tumor pathology data and breast tumor imaging data.
[0048] Preferably, the breast tumor pathology feature processing module includes the following functions:
[0049] Extracting pathological biomarker feature data based on breast tumor pathology data to generate pathological biomarker feature data;
[0050] Analyze the pathological type of breast tumors according to the characteristic data of pathological biomarkers to generate breast tumor pathological type data;
[0051] The breast tumor mitosis index is analyzed according to the pathological biomarker characteristic data to generate the breast tumor mitosis index.
[0052] The present invention extracts pathological biomarker feature data based on breast tumor pathological data. By analyzing the pathological data of breast tumors, a variety of important biomarker features are extracted, such as cell structure, nuclear morphology, cell distribution, etc., to provide detailed information about the biological characteristics and pathological characteristics of the tumor. The generated pathological biomarker feature data can not only be used for the diagnosis and classification of breast tumors, but also can be used as input for subsequent analysis and prediction models to help evaluate the development of tumors and the prognosis of patients. Breast tumor pathological type analysis is performed based on the pathological biomarker feature data. Based on the extracted biomarker feature data, the system can accurately classify the pathological type of breast tumors, such as benign tumors, malignant tumors, etc., which is crucial for correct pathological diagnosis and treatment planning. Different types of breast tumors require different treatment strategies. By accurately classifying tumor types, doctors can be provided with personalized treatment suggestions to improve treatment effects and patient survival rates. The mitosis index of breast tumors is analyzed based on the pathological biomarker characteristic data. The mitosis index is an important indicator for evaluating the proliferation rate of tumor cells. By analyzing the biomarker characteristic data, the mitosis index of breast tumors is calculated, which is of great significance for evaluating the growth rate of tumors, predicting their invasiveness and selecting appropriate treatment plans.
[0053] In an embodiment of the present invention, natural language processing (NLP) technology is used to analyze the text of the pathology report, and a named entity recognition (NER) algorithm is applied to identify key biomarkers, such as ER, PR, HER2, and Ki-67, and regular expressions or pattern matching are used to extract the specific values or states of these markers. For example: ER status: positive / negative, and positive rate (such as 70%); PR status: positive / negative, and positive rate; HER2 status: 0 / 1+ / 2+ / 3+ or negative / positive; Ki-67 proliferation index: percentage value. Other relevant pathological features, such as tumor size, grade, lymph node metastasis status, etc., are extracted to generate pathological biomarker feature data. Based on the extracted biomarker features, a rule-based classification system is established to define the judgment rules for different breast tumor pathological types according to the classification standards of WHO or other authoritative organizations. For example: if ER and / or PR are positive and HER2 is negative, it is defined as "Luminal A type"; if ER and / or PR are positive and HER2 is positive, it is defined as "Luminal B type"; if ER and PR are negative and HER2 is positive, it is defined as "HER2 overexpression type"; if ER, PR and HER2 are all negative, it is defined as "triple negative". According to the analyzed breast tumor pathological types, the breast tumor pathological type data is obtained. Search for keywords and phrases related to nuclear division in the pathology report text, such as "mitotic count", "mitoses per 10 HPF", etc., and use regular expressions to extract the specific value of nuclear division count. If the nuclear division index or level is directly given in the report, extract the information directly. If there is only raw count data, calculate the nuclear division index according to the standard. For example: Level 1: 0-9 divisions / 10 HPF; Level 2: 10-19 divisions / 10 HPF; Level 3: ≥20 divisions / 10 HPF. If mitotic count information is missing, try using the Ki-67 proliferation index as a surrogate marker, as they usually correlate well with the resulting breast tumor mitotic index.
[0054] Preferably, the breast tumor target area image extraction module includes the following functions:
[0055] Perform grayscale gradient feature analysis on breast tumor image data to generate grayscale gradient feature data;
[0056] Performing corner point target detection processing on breast tumor image data according to gray value gradient feature data to generate breast tumor target detection data;
[0057] The target region image segmentation processing is performed on the breast tumor image data according to the breast tumor target detection data to generate breast tumor target region image data.
[0058] The present invention performs grayscale gradient feature analysis on breast tumor image data. By performing grayscale gradient feature analysis on breast tumor image data, grayscale changes and gradient information in the image are extracted to help distinguish tumor areas from normal tissues, and improve the accuracy and precision of the target area. The grayscale gradient feature data reflects the subtle changes in the image, can capture the details and edge information of the tumor area, and provide more information support for subsequent target detection and segmentation. Corner point target detection processing is performed on breast tumor image data based on the grayscale gradient feature data, effectively locating the position and boundary of the breast tumor in the image, accurately marking the edge and key area of the tumor, and providing an accurate starting point for subsequent segmentation and analysis. Corner point target detection can automatically identify potential tumor areas, reduce manual intervention and the complexity of operation, and improve processing efficiency and consistency. The target area image segmentation processing of breast tumor image data is performed based on the breast tumor target detection data to accurately extract the outline and internal structure of the tumor, providing clear image data for further analysis and diagnosis. The generated breast tumor target area image data is used in clinical diagnosis and treatment planning, helping doctors to better understand the size, shape and location of the tumor, and providing support for the formulation of personalized treatment plans.
[0059] In the embodiment of the present invention, the grayscale gradient feature analysis is performed on the breast tumor image data, the cvtColor function of the OpenCV library is used to convert the input color breast image into a grayscale image, and then the GaussianBlur function is applied to perform Gaussian filtering (kernel size is set to 5x5, σ value is 1.5) to reduce noise, the Sobel function is used to calculate the gradients in the x and y directions respectively (ksize is set to 3), and the gradient amplitude and direction are calculated by the sqrt and arctan2 functions of the numpy library. Then the gradient amplitude histogram of the entire image is calculated (using the histogram function of numpy, bins is set to 256), and the image is divided into a 16x16 grid, and the hog function of the skimage library is used to calculate the HOG feature for each grid (orientations is set to 9, pixels_per_cell is set to (8,8), and cells_per_block is set to (2,2)). At the same time, the local_binary_pattern function of skimage is used to calculate the LBP feature of each grid (P value is set to 8, R value is set to 1). The global gradient histogram, HOG features and LBP features were connected into a feature vector using the concatenate function of numpy, and then the PCA class of the sklearn library was used for dimensionality reduction (n_components was set to 0.95, i.e., 95% of the variance information was retained). The MinMaxScaler class of sklearn was used to normalize the features after dimensionality reduction, and the values were scaled to the range of [0, 1] to obtain the final grayscale gradient feature data. Corner point target detection was performed on the breast tumor image data based on the grayscale gradient feature data, and Harris corner point detection was performed using the cornerHarris function of OpenCV (blockSize was set to 2, ksize was set to 3, and k was set to 0.04), and then significant corner points were screened by setting the threshold (the threshold value was 1% of the maximum value of the cornerHarris return result). Then use the cornerSubPix function to perform sub-pixel precision on the corner position (winSize is set to (5,5), zeroZone is set to (-1,-1), and criteria is set to (cv2.TERM_CRITERIA_EPS + cv2.TERM_CRITERIA_MAX_ITER, 40, 0.001)). Then use the BRIEF class to calculate the BRIEF descriptor of the 11x11 area around each corner point (bytes is set to 32).Use the FlannBasedMatcher class for feature matching (index_params is set to dict(algorithm = FLANN_INDEX_LSH, table_number = 6, key_size = 12, multi_probe_level = 1), search_params is set to dict(checks=50)). Use sklearn's DBSCAN class to cluster the matched corner points (eps is set to 20, min_samples is set to 5). For each cluster, use the cv2.minAreaRect function to calculate the minimum enclosing rectangle as the candidate target area. Use the previously generated grayscale gradient feature data and scipy's correlate2d function to calculate the correlation between each candidate area and the pre-trained tumor model as a score. Use the non-maximum suppression algorithm (for implementation, refer to tensorflow's tf.image.non_max_suppression function, iou_threshold is set to 0.5) to remove overlapping areas, and retain the candidate area information (including location, size, score, etc.) to obtain breast tumor target detection data. According to the breast tumor target detection data, the target region image segmentation processing is performed on the breast tumor image data. Based on the target detection data, the zeros function of numpy is used to create the initial segmentation mask, and then the cv2.snake function is used to implement the active contour algorithm for fine segmentation (alpha is set to 0.1, beta is set to 0.1, and gamma is set to 0.1). Then, the detected target center is used as the seed point to implement the region growing algorithm (the threshold is set to 20, that is, the new pixel is included in the region when the grayscale difference between the new pixel and the seed point is less than 20). The cv2.distanceTransform and cv2.watershed functions are used to implement the watershed algorithm for boundary optimization. The segmentation results are opened and closed using the cv2.morphologyEx function (kernel size is set to 5x5, iterations is set to 2) to smooth the boundaries. The majority voting method is implemented (the final category of each pixel is determined by the majority of the results of the three methods of active contour, region growing and watershed) to combine the results of different segmentation methods.The pydensecrf library was used to implement conditional random fields to optimize the spatial consistency of segmentation boundaries (pos_w was set to 3, pos_xy_std was set to 1, bi_w was set to 4, bi_xy_std was set to 67, and bi_rgb_std was set to 3). The morphological features (such as area, perimeter, roundness, etc.) and texture features of the segmented region were calculated (skimage's greycoprops function was used to calculate GLCM features, including contrast, correlation, energy, etc.). The h5py library was used to create an HDF5 file, in which the original image, segmentation mask, boundary coordinates, and extracted features were saved to obtain the final breast tumor target area imaging data.
[0060] Preferably, the breast tumor growth trend analysis module includes the following functions:
[0061] Performing breast tumor vascular fluid characteristic analysis based on breast tumor target area image data to generate breast tumor vascular fluid characteristic data;
[0062] Performing breast tumor metastasis characteristic analysis based on breast tumor vascular fluid characteristic data to generate breast tumor metastasis characteristic data;
[0063] Performing breast tumor metastasis pattern analysis and processing according to breast tumor metastasis characteristic data to generate breast tumor metastasis pattern data;
[0064] Breast tumor growth trend analysis is performed based on breast tumor nuclear division index and breast tumor metastasis pattern data to generate breast tumor growth trend data.
[0065] The present invention performs a breast tumor vascular fluid characteristic analysis based on the image data of the target area of the breast tumor, reveals the vascular structure and blood flow dynamics of the tumor, and is of great significance for evaluating the blood supply of the tumor, predicting the growth rate of the tumor and the prognosis. The generated vascular fluid characteristic data can provide detailed information on blood flow velocity, vascular density and vascular morphology, and help doctors understand the nutrition supply status and internal environment of the tumor. Performing a breast tumor metastasis characteristic analysis based on the breast tumor vascular fluid characteristic data to evaluate the metastasis tendency and possibility of the tumor is of great significance for judging the aggressiveness of the tumor and the choice of treatment strategy. The generated metastasis characteristic data can help doctors predict the metastasis potential of the tumor and provide support for the formulation of personalized treatment plans and prognosis evaluation. Performing a breast tumor metastasis pattern analysis and processing based on the breast tumor metastasis characteristic data, identifying and classifying different metastasis patterns, such as lymphatic metastasis, hematogenous metastasis, etc., understanding the metastasis path and diffusion mode of the tumor, and preventing or treating the metastasis of the tumor in a targeted manner, improving the treatment effect and the survival rate of the patient. Breast tumor growth trend analysis is performed based on the breast tumor mitosis index and breast tumor metastasis pattern data. Growth trend analysis combined with the mitosis index and metastasis pattern data can evaluate the growth rate and proliferation activity of breast tumors, provide detailed information on tumor growth and development, and support the scientific basis for prognostic evaluation and decision-making.
[0066] As an example of the present invention, refer to Figure 2 As shown, Figure 1 Schematic diagram of the functional flow of the breast tumor growth trend analysis module in FIG. 1 . In this example, the functions of the breast tumor growth trend analysis module include:
[0067] S401: Analyzing the characteristics of the blood vessels of the breast tumor based on the image data of the target area of the breast tumor to generate the blood vessels of the breast tumor.
[0068] In the embodiment of the present invention, the skimage.filters.frangi function of the scikit-image library is used to perform Frangi filtering on the target area image (scale_range is set to (1, 10), scale_step is set to 2, beta1 is set to 0.5, and beta2 is set to 15) to enhance the vascular structure, the skimage.morphology.skeletonize function is used to perform skeleton processing on the enhanced image, the vascular centerline is extracted, the networkx library is used to construct a vascular network graph, the vascular centerline is converted into a graph structure (nodes represent vascular branch points, and edges represent vascular segments), and the topological features of the network, such as average degree centrality (nx.degree_centrality), betweenness centrality (nx.betweenness_centrality), and closeness centrality (nx.closeness_centrality), are calculated, the scipy.ndimage.measurements.label function is used to perform connected domain analysis, and the vascular density (ratio of the number of vascular pixels to the total number of pixels) and branch complexity (ratio of the number of endpoints to the number of branch points) are calculated, and the skimage.measure.regionprops function is used to calculate the morphological features of the blood vessels, such as curvature, tortuosity, etc. The cv2.calcOpticalFlowFarneback function of OpenCV was used to estimate the optical flow field between two consecutive frames to analyze the hemodynamic characteristics, and all the extracted features (including topological features, density, complexity, morphological features, and dynamic features) were combined to obtain the breast tumor vascular fluid characteristic data.
[0069] S402: Performing breast tumor metastasis characteristic analysis based on breast tumor vascular fluid characteristic data to generate breast tumor metastasis characteristic data;
[0070] In the embodiment of the present invention, sklearn.cluster.KMeans is used for cluster analysis (n_clusters is set to 3, and vascular features are divided into three categories of low, medium and high risk based on experience) to identify potential metastasis risk patterns, and sklearn.ensemble.RandomForestClassifier is used to build a random forest model (n_estimators is set to 100, max_depth is set to 10), and vascular features are associated with known metastasis states to obtain feature importance rankings, and sklearn.metrics.roc_auc_score is used to calculate the ROC-AUC value of each feature to evaluate its contribution to metastasis prediction. The scipy.stats module is used to perform hypothesis tests (such as t-tests or Mann-Whitney U tests) to compare the significance of differences in each feature between metastasis and non-metastasis groups. Radiomics feature extraction is implemented, and the pyradiomics library is used to calculate first-order statistics, shape features, texture features, etc. sklearn.feature_selection.RFE was applied to perform recursive feature elimination and select the most relevant feature subset. The screened features, their importance scores, and statistical test results were used to obtain breast tumor metastasis feature data.
[0071] S403: performing breast tumor metastasis pattern analysis and processing according to the breast tumor metastasis characteristic data to generate breast tumor metastasis pattern data;
[0072] In the embodiment of the present invention, sklearn.manifold.TSNE is applied to perform t-SNE dimensionality reduction (n_components is set to 2, perplexity is set to 30) to visualize the high-dimensional feature distribution. sklearn.cluster.DBSCAN is used for density clustering (eps is set to 0.5, min_samples is set to 5) to identify different transfer mode groups, and networkx is used to build a feature association network. The edge weight is based on the mutual information between features (calculated using sklearn.metrics.mutual_info_score). The networkx.algorithms.community.greedy_modularity_communities function is used to perform community detection and discover feature modules. Implement a deep learning model based on graph convolutional networks (using the pytorch_geometric library) to capture the complex nonlinear relationships between features, use the lifelines library to implement the Cox proportional hazard model, analyze the impact of different feature combinations on metastasis time, apply sklearn.tree.DecisionTreeClassifier to build a decision tree model (max_depth is set to 5), extract interpretable metastasis rules, use the SHAP (SHapley Additive exPlanations) library to calculate the SHAP value of the features, explain the model prediction, implement the Markov chain Monte Carlo (MCMC) method, perform Bayesian inference, estimate the posterior distribution of the metastasis probability, and combine the clustering results, network analysis results, survival analysis results, decision rules, and SHAP values to obtain breast tumor metastasis pattern data.
[0073] S404: Analyze the growth trend of the breast tumor according to the breast tumor nuclear division index and the breast tumor metastasis pattern data to generate breast tumor growth trend data.
[0074] In an embodiment of the present invention, breast tumor growth trend analysis is performed based on breast tumor nuclear division index and breast tumor metastasis pattern data, scipy.stats.pearsonr is used to calculate the correlation coefficient between nuclear division index and each metastasis feature, and the statsmodels library is used to implement the generalized additive model (GAM), analyze the nonlinear effect of nuclear division index and metastasis feature on tumor volume, implement the dynamic time warping (DTW) algorithm, and compare the similarity of tumor growth curves of different patients. The prophet library is used to construct a time series prediction model to predict the change of tumor volume in the short term, and the gradient boosting regression tree is implemented using sklearn.ensemble.GradientBoostingRegressor to predict the long-term tumor growth trend. A cellular automaton model is implemented to simulate the tumor growth process considering spatial constraints. Scipy.integrate.odeint is used to solve a group of ordinary differential equations to implement a mathematical model of tumor growth based on the Gompertz equation, and the PyMC3 library is used to implement the Bayesian hierarchical model to estimate the growth parameters at the population level and the individual level. Survival analysis technology (using the KaplanMeierFitter class of the lifelines library) is used to estimate the survival curves under different growth modes. Implement Monte Carlo simulation (using numpy.random module) to generate a large number of possible growth trajectories, evaluate the uncertainty of prediction, and combine prediction results, model parameters, survival curves, uncertainty estimates, etc. to obtain breast tumor growth trend data.
[0075] Preferably, the breast tumor vascular fluid characteristic analysis based on the breast tumor target area image data includes:
[0076] Performing target area blood vessel density analysis on breast tumor target area image data to generate breast tumor target area blood vessel density data;
[0077] Performing target region fluid characteristic analysis on breast tumor target region image data to generate breast tumor target region fluid characteristic data;
[0078] Breast tumor vascular fluid characteristic analysis is performed based on the breast tumor target area vascular density data and the breast tumor target area fluid characteristic data to generate breast tumor vascular fluid characteristic data.
[0079] The present invention performs target area vascular density analysis on target area image data of breast tumors to evaluate the blood supply of tumors. High vascular density is usually related to sufficient nutrition supply and strong invasiveness of tumors. The generated vascular density data can be used as a standard for evaluating the severity and spread of breast tumor lesions. Performing target area fluid characteristic analysis on target area image data of breast tumors can reveal the distribution and movement of fluid inside the tumor, understand the internal structure and fluid composition of the tumor, and provide an important basis for clinical diagnosis and treatment strategies. The generated fluid characteristic data can reflect the dynamic changes of fluid inside the tumor, including the accumulation, diffusion and discharge of fluid, which is helpful to monitor the growth of tumors and the effect of response treatment. Performing breast tumor vascular fluid characteristic analysis based on breast tumor target area vascular density data and breast tumor target area fluid characteristic data can more comprehensively understand the vascular system and fluid dynamic characteristics of breast tumors. The generated vascular fluid characteristic data not only assists in diagnosing the nature and characteristics of tumors, but also provides a scientific basis for formulating personalized treatment strategies, optimizing treatment effects and patient prognosis.
[0080] In the embodiment of the present invention, the target area blood vessel density analysis is performed on the target area image data of the breast tumor, the cv2.findContours function is used to detect the contour, and the cv2.drawContours function is used to draw the blood vessel contour, and the blood vessel area ratio (the number of blood vessel pixels divided by the total number of pixels) is calculated as the overall blood vessel density, and the skimage.measure.label function is used to mark the blood vessels, and then the skimage.measure.regionprops function is used to calculate the properties of each blood vessel area (such as area, perimeter, eccentricity, etc.), and the numpy histogram function is used to calculate the blood vessel size distribution histogram, and the scipy.ndimage.distance_transform_edt function is used to calculate the distance from each non-vascular pixel to the nearest blood vessel to generate a distance map. Based on the distance map, the mean and std functions of numpy are used to calculate the average blood vessel spacing and its standard deviation. Finally, the calculated indicators (overall density, size distribution, average spacing, etc.) are encapsulated using the DataFrame of the pandas library to obtain the blood vessel density data of the target area of the breast tumor. The target area fluid characteristics of breast tumor target area image data were analyzed, and the optical flow field between two consecutive frames was calculated using OpenCV's cv2.calcOpticalFlowFarneback function (pyr_scale was set to 0.5, levels was set to 3, winsize was set to 15, iterations was set to 3, poly_n was set to 5, and poly_sigma was set to 1.2). The amplitude of the optical flow field was calculated using numpy's hypot function, the mean function calculated the average flow velocity, and the std function calculated the standard deviation of the flow velocity. The scipy.stats module was used to calculate the deviation of the flow velocity distribution. The skew and kurtosis of the flow field are calculated using the gradient function of numpy, and then the divergence and curl are calculated. The flow field is clustered using the sklearn.cluster.KMeans function (n_clusters is set to 3) to identify different flow patterns. The power spectral density analysis is performed using the scipy.signal.welch function to identify the main frequency components of the flow field. The entropy of the flow field is calculated using the skimage.measure.shannon_entropy function to evaluate the complexity of the flow.The analysis based on particle image velocimetry (PIV) was implemented. The process_pair function of the openpiv library (window_size set to 32, overlap set to 16) was used to calculate the local velocity vector. All extracted features (average velocity, standard deviation, skewness, kurtosis, divergence, curl, clustering results, spectrum characteristics, entropy, PIV results, etc.) were encapsulated using pandas's DataFrame to obtain the fluid characteristic data of the target area of breast tumors. The blood vessel fluid characteristics of breast tumors were analyzed based on the blood vessel density data of the target area of breast tumors and the fluid characteristic data of the target area of breast tumors. The two sets of features were merged using the concatenate function of numpy. sklearn.decomposition.PCA was applied for principal component analysis (n_components set to 0.95) to reduce the feature dimension. The correlation coefficient matrix between the blood vessel density features and the fluid features was calculated using scipy.stats.pearsonr. The feature association network was constructed using the networkx library, and the edge weights were based on the absolute value of the correlation coefficient. The networkx.algorithms.centrality module was used to calculate the centrality index of each node in the network (such as degree centrality, betweenness centrality, and eigenvector centrality). The sklearn.cluster.AgglomerativeClustering was used for hierarchical clustering (n_clusters was set to 4) to identify feature combination patterns. The sklearn.ensemble.RandomForestRegressor was used to build a random forest model (n_estimators was set to 100 and max_depth was set to 10). The influence of features on tumor growth was analyzed. The shap library was used to calculate the SHAP (SHapley Additive exPlanations) value and interpret the model prediction. The scipy.stats module was used for hypothesis testing (such as t-test or Mann-Whitney U test). The differences in vascular fluid characteristics of different subtypes of tumors were compared. A Bayesian network model based on PyMC3 was implemented to infer the causal relationship between features. The analysis results (including features after dimensionality reduction, correlation coefficient matrix, network centrality index, clustering results, feature importance, SHAP value, statistical test results, causal relationship diagram, etc.) were used as the final breast tumor vascular fluid characteristic data.
[0081] Preferably, the breast tumor metastasis pattern analysis and processing according to the breast tumor metastasis characteristic data comprises:
[0082] Performing breast tumor metastasis characteristic clustering analysis on breast tumor metastasis characteristic data based on a preset clustering algorithm to generate clustered breast tumor metastasis characteristic data;
[0083] Breast tumor metastasis pattern analysis and processing are performed based on clustered breast tumor metastasis feature data to generate breast tumor metastasis pattern data.
[0084] The present invention performs cluster analysis of breast tumor metastasis characteristic data based on a preset clustering algorithm. Through cluster analysis, breast tumor cases with similar metastasis characteristics are grouped into clusters, and different metastasis patterns, such as lymphatic metastasis, hematogenous metastasis, etc., are identified and distinguished. The complex metastasis characteristic data are reduced to fewer and more representative clusters, simplifying the data structure and helping doctors better understand the diversity of metastasis patterns of breast tumors. Breast tumor metastasis pattern analysis and processing are performed based on clustered breast tumor metastasis characteristic data. Based on the clustered metastasis characteristic data, the breast tumor metastasis pattern in each cluster is deeply analyzed, including the type, frequency and path of metastasis, providing a clearer metastasis trend assessment, helping to predict the development of tumors and the prognosis of patients. The generated metastasis pattern data provides a basis for formulating personalized treatment strategies. According to the characteristics of different metastasis patterns, the most appropriate treatment plan and management strategy are selected to optimize the treatment effect and survival rate of patients.
[0085] In the embodiment of the present invention, a breast tumor metastasis feature clustering analysis is performed on the breast tumor metastasis feature data based on a preset clustering algorithm, and sklearn.cluster.KMeans is used as the preset clustering algorithm (n_clusters is set to 5, and the optimal number of clusters is determined based on experience or by the silhouette coefficient) to cluster the breast tumor metastasis feature data. The sklearn.metrics module is used to calculate clustering evaluation indicators, such as the silhouette coefficient (silhouette_score) and the Calinski-Harabasz index (calinski_harabasz_score), and sklearn.manifold.TSNE is applied to perform t-SNE dimensionality reduction (n_components is set to 2, perplexity is set to 30), and the center point (center of mass) of each cluster and the mean and standard deviation of each feature are calculated. The scipy.stats module was used to perform hypothesis tests (such as ANOVA or Kruskal-Wallis test) on each cluster, identify features that significantly distinguish clusters, and use cluster labels, features after dimensionality reduction, cluster evaluation indicators, cluster centers, feature statistics, and significance test results as cluster breast tumor metastasis feature data. Breast tumor metastasis pattern analysis was performed based on cluster breast tumor metastasis feature data. sklearn.tree.DecisionTreeClassifier was used to build a decision tree model (max_depth was set to 5) to extract interpretable metastasis rules. sklearn.ensemble.RandomForestClassifier was used to build a random forest model (n_estimators was set to 100, max_depth was set to 10) to analyze feature importance. The shap library was used to calculate SHAP values and explain the model's predictions for each cluster. The lifelines library was used to implement the Cox proportional hazard model to analyze the survival characteristics of different clusters. Use networkx to build a feature association network, and the edge weights are based on the mutual information between features (calculated using sklearn.metrics.mutual_info_score). Use the networkx.algorithms.community.greedy_modularity_communities function to perform community detection and discover feature modules. Use the scipy.stats module to perform hypothesis testing, compare the differences in various clinical indicators between different clusters, implement the Markov Chain Monte Carlo (MCMC) method for Bayesian inference, estimate the transfer probability distribution of each cluster, and obtain transfer rules, feature importance, SHAP values, survival analysis results, network analysis results, model performance evaluation, statistical test results, and probability distribution, thereby generating breast tumor metastasis pattern data.
[0086] Preferably, the breast tumor growth trend analysis based on the breast tumor nuclear division index and breast tumor metastasis pattern data includes:
[0087] Performing metastasis pattern time series feature analysis on breast tumor metastasis pattern data to generate metastasis pattern time series feature data;
[0088] Perform preliminary analysis on breast tumor growth trend based on the temporal characteristic data of the metastasis pattern to generate preliminary breast tumor growth trend data;
[0089] Perform breast tumor metastasis survival relationship analysis based on metastasis pattern time series feature data to generate breast tumor metastasis survival relationship data;
[0090] According to the breast tumor metastasis survival relationship data and the breast tumor nuclear division index, the breast tumor nonlinear growth factor analysis is performed to generate the breast tumor nonlinear growth factor;
[0091] Based on the nonlinear growth factor of the breast tumor, the preliminary breast tumor growth trend data is optimized and adjusted for breast tumor growth trend accuracy to obtain breast tumor growth trend data.
[0092] The present invention performs a transfer pattern time series feature analysis on breast tumor metastasis pattern data to reveal the dynamic changes in the transfer process, which is used to assist in understanding the development trend and pattern changes of tumor metastasis, and provide basic data for subsequent growth trend analysis. The generated time series feature data shows the time series changes of the breast tumor metastasis pattern to identify time-dependent factors related to the growth trend. A preliminary analysis of the growth trend of the breast tumor is performed based on the transfer pattern time series feature data. The preliminary analysis of the growth trend of the breast tumor can predict the growth rate and preliminary development direction of the tumor, which provides a basis for the formulation of treatment plans and helps early intervention and treatment optimization. A breast tumor metastasis survival relationship analysis is performed based on the transfer pattern time series feature data. Analysis of the relationship between breast tumor metastasis pattern and survival can help evaluate the patient's prognosis. The generated survival relationship data reveals the impact of different transfer patterns on the patient's survival period. Nonlinear growth factor analysis of breast tumors is performed based on breast tumor metastasis-survival relationship data and breast tumor nuclear division index. Analyzing nonlinear growth factors can help understand the dynamic process of breast tumor growth, including changes in growth rate and possible growth trends. As an important reference for predicting the speed and behavior of tumor development, the generated nonlinear growth factor data can help develop personalized treatment strategies and adjust treatment plans according to the dynamic changes of tumor growth to improve treatment effects and patient survival rates. Based on the nonlinear growth factor of breast tumors, the initial breast tumor growth trend data is optimized and adjusted for breast tumor growth trend accuracy. Optimizing growth trend data in combination with nonlinear growth factors can improve the accuracy and predictive ability of growth trend analysis, more accurately understand the growth pattern and development trend of breast tumors, and provide a more reliable basis for treatment decisions.
[0093] In an embodiment of the present invention, a transfer pattern time series feature analysis is performed on the breast tumor metastasis pattern data, and a time series analysis tool in the statsmodels library, such as an ARIMA model, is used to extract and analyze the time series features of the breast tumor metastasis pattern data. By calculating the moving average trend analysis method, specific transfer pattern time series feature data is generated, and scipy.stats is used in combination to perform data distribution analysis to ensure the statistical significance and accuracy of the time series feature data. A preliminary analysis of the growth trend of the breast tumor is performed based on the transfer pattern time series feature data. Based on the generated transfer pattern time series feature data, a preliminary analysis of the growth trend of the breast tumor is achieved by applying a generalized additive model (GAM). The GAM tool in the statsmodels library of Python is used, combined with the transfer pattern time series feature data of the breast tumor, a nonlinear model is established to describe the influence of the nuclear division index and each transfer feature on the tumor volume, and the GAM model is used to select variables and optimize parameters to generate preliminary breast tumor growth trend data, including predicted tumor volume changes and related uncertainty analysis. According to the metastasis pattern time series characteristic data, the breast tumor metastasis survival relationship analysis is performed. Survival analysis techniques, such as the KaplanMeierFitter class in Python's lifelines library, are used to analyze the metastasis pattern time series characteristic data. The Kaplan-Meier method is implemented to estimate the patient survival function. The effect of breast tumor metastasis pattern on survival time is considered. The survival difference between different metastasis patterns is evaluated through the Log-rank test or the Cox proportional hazard model to generate breast tumor metastasis survival relationship data. According to the breast tumor metastasis survival relationship data and the breast tumor nuclear division index, the breast tumor nonlinear growth factor analysis is performed. Combined with the breast tumor metastasis survival relationship data and the nuclear division index, the odeint function in Python's scipy.integrate library is used to realize the tumor growth mathematical model based on the Gompertz equation. By solving the ordinary differential equation group, the breast tumor nonlinear growth factor, that is, the growth parameter of the Gompertz model, is estimated, which reflects the growth dynamics of breast tumors under different metastasis patterns and nuclear division index conditions, and provides a basis for subsequent growth trend data analysis.Based on the nonlinear growth factor of breast tumor, the accuracy of breast tumor growth trend of preliminary breast tumor growth trend data was optimized and adjusted. The prophet library in Python was used to build a time series prediction model, and the nonlinear growth factor of breast tumor was used as input data for trend analysis to predict the change of breast tumor volume in the short term. The GradientBoostingRegressor in the sklearn.ensemble library was used to implement the gradient boosting regression tree to predict the long-term tumor growth trend. The uncertainty of the prediction was evaluated by Monte Carlo simulation. The final breast tumor growth trend data was generated by optimizing the accuracy of the preliminary breast tumor growth trend data.
[0094] Preferably, the prognostic breast tumor resection risk prediction module includes the following functions:
[0095] Acquire historical breast tumor cutting data from a medical database, wherein the historical breast tumor cutting data includes historical breast tumor cutting area data and historical cutting risk label data;
[0096] Based on the preset convolutional neural network algorithm and historical breast tumor cutting data, a relationship prediction model between the breast tumor cutting area and the cutting risk is established to construct a breast tumor cutting risk prediction model;
[0097] Performing prognostic breast tumor target area image analysis on breast tumor target area image data according to breast tumor growth trend data to generate prognostic breast tumor target area image data;
[0098] The prognostic breast tumor target area image data is transmitted to the breast tumor cutting risk prediction model for prognostic breast tumor cutting risk prediction, thereby generating prognostic breast tumor cutting risk prediction data.
[0099] The present invention obtains historical breast tumor cutting data from a medical database to accumulate a large number of actual surgical records, provides a rich data basis for model establishment, helps evaluate the prediction accuracy of cutting risk and the adaptability of the model, and provides a labeled training sample for the model, which helps to establish an accurate cutting risk prediction model and improve the credibility and accuracy of the prediction. A relationship prediction model between a breast tumor cutting area and cutting risk is established based on a preset convolutional neural network algorithm and historical breast tumor cutting data. The convolutional neural network algorithm can effectively learn the complex relationship between the breast tumor cutting area and the cutting risk, process the spatial relationship and features in the image data, and improve the prediction accuracy of the cutting risk. According to the breast tumor growth trend data, the breast tumor target area image data is analyzed for prognosis of the breast tumor target area image data. The image analysis combined with the growth trend data can update the patient's prognosis of the breast tumor target area image data in real time, more accurately reflect the patient's tumor status and development trend, and assist in adjusting the treatment strategy, such as planning the cutting decision of the breast tumor. Predicting the risk in advance can prompt the necessary intervention measures to be taken before the disease worsens, which helps to improve the patient's treatment success rate and survival rate. Transmitting the prognostic breast tumor target area imaging data to the breast tumor cutting risk prediction model for prognostic breast tumor cutting risk prediction can generate breast tumor cutting risk prediction results for individual patients, provide personalized surgical risk assessment and decision support, and reduce surgical risks and the possibility of complications.
[0100] As an example of the present invention, refer to Figure 2 As shown, Figure 1 The functional flow chart of the prognostic breast tumor resection risk prediction module in this example includes:
[0101] S501: Acquire historical breast tumor cutting data from a medical database, wherein the historical breast tumor cutting data includes historical breast tumor cutting area data and historical cutting risk label data;
[0102] In an embodiment of the present invention, historical breast tumor cutting data in a medical database, including breast tumor cutting area data and cutting risk label data, is obtained, and the pandas library in Python is used to load and clean the data to ensure data quality. For breast tumor cutting area data, it is necessary to parse the image file or use a medical image processing library (such as pydicom) to read DICOM format image data. At the same time, the historical cutting risk label data is sorted, including risk level classification or probability distribution. To ensure data integrity and privacy security, appropriate data encryption and access control measures are adopted.
[0103] S502: Establishing a relationship prediction model between a breast tumor cutting area and cutting risk based on a preset convolutional neural network algorithm and historical breast tumor cutting data, so as to construct a breast tumor cutting risk prediction model;
[0104] In an embodiment of the present invention, a convolutional neural network (CNN) algorithm is used to establish a relationship prediction model between the cutting area of a breast tumor and the cutting risk, and a CNN model is constructed using the TensorFlow or PyTorch framework in Python. Design a suitable network structure, such as a multi-layer convolution layer, a pooling layer, and a fully connected layer. In the model training stage, the historical breast tumor cutting data is divided into a training set and a validation set. The model performance is evaluated by a cross-validation or retention set method. The model training is performed using an appropriate loss function (such as a cross entropy loss), an optimizer (such as an Adam optimizer), and a learning rate scheduling strategy. The training process is accelerated using a GPU to speed up the calculation speed and improve the training efficiency. According to the trained CNN model, the cutting area and cutting risk of a new breast tumor cutting image are predicted. The reliability and accuracy of the model in actual clinical applications are improved by interpretative analysis and uncertainty assessment of the prediction results. The performance indicators of the model are monitored in real time, and the model is adjusted and retrained when necessary to ensure its continued effectiveness and adaptability.
[0105] S503: performing prognostic breast tumor target area image analysis on the breast tumor target area image data according to the breast tumor growth trend data to generate prognostic breast tumor target area image data;
[0106] In the embodiment of the present invention, the active_contour function of the skimage.segmentation module is used to implement dynamic contour segmentation and extract tumor boundaries. Image matrix operations are implemented based on the numpy library, and the segmentation results are enlarged or deformed according to the breast tumor growth trend data. The cv2.resize function of the opencv-python library is used to adjust the image size to ensure consistency with the model input size, and the processed image data is converted into a numpy array, and the save function of numpy is used to save it in .npy format as the prognosis breast tumor target area image data.
[0107] S504: Transmitting the prognostic breast tumor target area image data to the breast tumor cutting risk prediction model to perform prognostic breast tumor cutting risk prediction, and generating prognostic breast tumor cutting risk prediction data.
[0108] In an embodiment of the present invention, the preprocessed prognostic breast tumor target area image data is loaded using numpy's load function, the torch.from_numpy function is used to convert the numpy array into a PyTorch tensor, and .to(device) is used to move the data to the GPU (if available). Load the previously saved breast tumor resection risk prediction model parameters, set the model to evaluation mode (model.eval()), and use the torch.no_grad() context manager to disable gradient calculation. The prediction results are obtained by forward propagation of the model. Use torch.argmax to obtain the predicted category, and torch.softmax calculates the probability distribution of each category. Convert the prediction results back to a numpy array, and use pandas's DataFrame to encapsulate the predicted risk category, probability and other information to generate prognostic breast tumor resection risk prediction data, including the type of breast tumor resection surgery and the corresponding resection risk probability, etc.
[0109] Preferably, the establishment of a relationship prediction model between breast tumor cutting area and cutting risk based on a preset convolutional neural network algorithm and historical breast tumor cutting data includes:
[0110] Using historical breast tumor cutting area data as input data and historical cutting risk label data as output data to design a breast tumor cutting risk training set;
[0111] Based on the preset convolutional neural network algorithm, a mapping relationship between breast tumor cutting area and cutting risk is established to construct a breast tumor cutting risk prediction model architecture;
[0112] The breast tumor cutting risk training set is transferred to the breast tumor cutting risk prediction model architecture for model training to obtain a breast tumor cutting risk prediction model.
[0113] The present invention uses historical breast tumor cutting area data as input data and historical cutting risk label data as output data to establish a correlation between the two. This correlation helps the training model understand the corresponding relationship between different cutting area features and cutting risks, thereby improving the accuracy and reliability of the prediction model. The historical breast tumor cutting area data as a training set can provide the model with rich samples, covering the cutting risk performance under a variety of different situations, which helps the model to learn and predict various potential cutting risk situations more comprehensively. The mapping relationship between breast tumor cutting area and cutting risk is established based on the preset convolutional neural network algorithm. The convolutional neural network can effectively capture the complex nonlinear mapping relationship between breast tumor cutting area and cutting risk, so that the model can extract features from high-dimensional image data and effectively associate these features with risk labels, thereby improving the accuracy and reliability of the prediction. The breast tumor cutting risk training set is transmitted to the breast tumor cutting risk prediction model architecture for model training. The trained model can process new breast tumor image data in real time, and quickly and accurately predict the patient's cutting risk, assisting in making more accurate decisions before surgery, and improving the success rate of surgery and the safety of patients.
[0114] In an embodiment of the present invention, the pandas library is used to read the previously saved historical breast tumor cutting area data and historical cutting risk label data, and the cutting risk label is converted to a numerical type using the LabelEncoder in sklearn.preprocessing. The train_test_split function of the sklearn.model_selection module is used to divide the data set into a training set, a validation set, and a test set in a ratio of 8:1:1, and the image data is enhanced. The albumentations library is used to implement random rotation, flipping, scaling, and other operations to increase data diversity. Create a custom Dataset class, inherit from torch.utils.data.Dataset, implement the __getitem__ and __len__ methods, and design a breast tumor cutting risk training set. The model architecture uses ResNet50 as the basic network, loads the pre-trained weights through torchvision.models, modifies the last fully connected layer of ResNet50, matches its output with the number of cutting risk categories, adds an attention mechanism, implements the SE (Squeeze-and-Excitation) module, enhances the model's attention to important features, implements the FPN (Feature Pyramid Network) structure, and extracts multi-scale features. Add several custom convolutional layers and fully connected layers after FPN for risk prediction. Use nn.Dropout to implement regularization to prevent overfitting. Choose an appropriate loss function, such as nn.CrossEntropyLoss for multi-classification problems, or nn.BCEWithLogitsLoss for multi-label classification, define an optimizer, such as torch.optim.Adam, and use the learning rate scheduler torch.optim.lr_scheduler.ReduceLROnPlateau to dynamically adjust the learning rate. Use the torchsummary library to print the model structure and check the number of parameters and computational complexity to build the breast tumor resection risk prediction model architecture. Load the preprocessed training and validation set data. Define the training function, which includes forward propagation, loss calculation, backpropagation, and parameter update steps. Implement the validation function to evaluate the model performance on the validation set. Use the tqdm library to create a progress bar to visualize the training process. In the training loop, each epoch traverses the entire training set and updates the model parameters after each mini-batch. Evaluate the model on the validation set regularly and record indicators such as loss and accuracy. Use tensorboard to record various indicators during the training process for easy visual analysis and to implement an early stopping mechanism. When the performance of the validation set no longer improves, training is stopped to prevent overfitting.After training, the model performance is evaluated on the test set, and the accuracy, precision, recall, and F1 score are calculated. The confusion matrix and ROC curve are generated using the sklearn.metrics module to comprehensively evaluate the model performance and obtain a breast tumor resection risk prediction model. The model is converted to TorchScript format using torch.jit.script for deployment in a production environment.
[0115] The beneficial effect of the present application is that the data intelligent analysis system before breast tumor resection of the present invention can accurately identify and extract the target area of breast tumor through the target area segmentation and feature extraction technology of breast tumor image data, and provide necessary data support for subsequent growth trend analysis and treatment decision-making. And the characteristics of nuclear division index and metastasis pattern of multi-source breast tumors are comprehensively analyzed, and the growth rate and diffusion tendency of tumors can be evaluated, and predictive information is provided to help adjust treatment strategies and prevent postoperative recurrence. According to breast tumor growth trend data and image analysis results, surgical risks are quantified and decision support is provided, so as to achieve data-driven treatment options. Through the prognosis breast tumor cutting risk prediction data and the prognosis breast tumor treatment auxiliary decision feedback personalized prognosis breast tumor resection information, it helps doctors optimize surgical plans and manage patients' postoperative recovery, and improves the patient's recovery status in the later stage. In order to realize the unified quantitative analysis of prognosis breast tumors in combination with multi-source breast tumor case information of patients, the image characteristics and pathological characteristics of the comprehensive breast tumor case information are better for the analysis of the morphology and pathological conditions of prognosis breast tumors, thereby providing accurate and efficient prognosis breast tumor treatment auxiliary decision-making.
[0116] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.
[0117] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.
Claims
1. A data intelligent analysis system before breast tumor resection, characterized by: Includes the following modules: A breast tumor case data preprocessing module is used to obtain breast tumor case data from a medical database; preprocess the breast tumor case data to obtain standard breast tumor case data, wherein the standard breast tumor case data includes breast tumor pathology data and breast tumor imaging data; A breast tumor pathology feature processing module is used to analyze the breast tumor pathology type based on the breast tumor pathology data and generate breast tumor pathology type data; and to calculate the breast tumor mitosis index based on the breast tumor pathology data and generate the breast tumor mitosis index; A breast tumor target area image extraction module is used to perform target area image segmentation processing on breast tumor image data to generate breast tumor target area image data; A breast tumor growth trend analysis module is used to analyze and process breast tumor metastasis patterns based on breast tumor target area imaging data to generate breast tumor metastasis pattern data; and to analyze breast tumor growth trends based on breast tumor nuclear division index and breast tumor metastasis pattern data to generate breast tumor growth trend data; A prognostic breast tumor resection risk prediction module is used to construct a breast tumor resection risk prediction model; perform prognostic breast tumor target area image analysis on breast tumor target area image data based on breast tumor growth trend data to generate prognostic breast tumor target area image data; transmit the prognostic breast tumor target area image data to the breast tumor resection risk prediction model to perform prognostic breast tumor resection risk prediction to generate prognostic breast tumor resection risk prediction data; The prognostic breast tumor resection information intelligent feedback module is used to analyze the prognostic breast tumor treatment characteristics based on the breast tumor pathological type data and the prognostic breast tumor target area imaging data to obtain the prognostic breast tumor treatment characteristic data; A prognostic breast tumor treatment auxiliary decision is designed based on the prognostic breast tumor treatment characteristic data; the prognostic breast tumor resection risk prediction data and the prognostic breast tumor treatment auxiliary decision are transmitted to the terminal to perform the prognostic breast tumor auxiliary resection information intelligent feedback operation.
2. The intelligent data analysis system before breast tumor resection according to claim 1, characterized in that: The breast tumor case data preprocessing module includes the following functions: Obtain breast tumor case data from medical databases; Performing breast tumor case information type analysis on breast tumor case data to obtain breast tumor case information type data; Design classification case information preprocessing decisions based on breast tumor case information type data; The breast tumor case data are preprocessed according to the classified case information preprocessing decision to obtain standard breast tumor case data.
3. The intelligent data analysis system before breast tumor resection according to claim 1, characterized in that: The breast tumor pathology feature processing module includes the following functions: Extracting pathological biomarker feature data based on breast tumor pathology data to generate pathological biomarker feature data; Analyze the pathological type of breast tumors based on the pathological biomarker characteristic data to generate breast tumor pathological type data; Breast tumor mitosis index analysis is performed based on pathological biomarker feature data to generate a breast tumor mitosis index.
4. The intelligent data analysis system before breast tumor resection according to claim 1, characterized in that: The breast tumor target area image extraction module includes the following functions: Perform grayscale gradient feature analysis on breast tumor image data to generate grayscale gradient feature data; Performing corner target detection processing on breast tumor image data based on grayscale gradient feature data to generate breast tumor target detection data; The target region image segmentation processing is performed on the breast tumor image data according to the breast tumor target detection data to generate breast tumor target region image data.
5. The intelligent data analysis system before breast tumor resection according to claim 1, characterized in that: The breast tumor growth trend analysis module includes the following functions: Analyze the fluid characteristics of breast tumor blood vessels based on the image data of the target area of the breast tumor to generate the fluid characteristics data of the breast tumor blood vessels; Performing breast tumor metastasis feature analysis based on breast tumor vascular fluid feature data to generate breast tumor metastasis feature data; Performing breast tumor metastasis pattern analysis and processing based on breast tumor metastasis characteristic data to generate breast tumor metastasis pattern data; Breast tumor growth trend analysis is performed based on breast tumor nuclear division index and breast tumor metastasis pattern data to generate breast tumor growth trend data.
6. The intelligent data analysis system before breast tumor resection according to claim 5, characterized in that: The breast tumor vascular fluid characteristic analysis based on the breast tumor target area image data includes: Performing target area blood vessel density analysis on breast tumor target area image data to generate breast tumor target area blood vessel density data; Performing target region fluid feature analysis on breast tumor target region image data to generate breast tumor target region fluid feature data; Breast tumor blood vessel fluid characteristic analysis is performed based on the blood vessel density data and fluid characteristic data of the breast tumor target area to generate breast tumor blood vessel fluid characteristic data.
7. The intelligent data analysis system before breast tumor resection according to claim 5, characterized in that: The breast tumor metastasis pattern analysis process according to the breast tumor metastasis characteristic data comprises: Performing breast tumor metastasis characteristic clustering analysis on the breast tumor metastasis characteristic data based on a preset clustering algorithm to generate clustered breast tumor metastasis characteristic data; Breast tumor metastasis pattern analysis and processing are performed based on clustered breast tumor metastasis feature data to generate breast tumor metastasis pattern data.
8. The intelligent data analysis system before breast tumor resection according to claim 5, characterized in that: The breast tumor growth trend analysis based on the breast tumor mitotic index and breast tumor metastasis pattern data includes: Performing metastasis pattern time series feature analysis on breast tumor metastasis pattern data to generate metastasis pattern time series feature data; Conduct preliminary analysis of breast tumor growth trends based on metastasis pattern temporal feature data to generate preliminary breast tumor growth trend data; Analyze the survival relationship of breast tumor metastasis based on the temporal characteristic data of metastasis pattern to generate breast tumor metastasis survival relationship data; Analyze the nonlinear growth factor of breast tumors based on the breast tumor metastasis-survival relationship data and the breast tumor nuclear division index to generate the nonlinear growth factor of breast tumors; The breast tumor growth trend accuracy of the preliminary breast tumor growth trend data is optimized and adjusted based on the breast tumor nonlinear growth factor to obtain the breast tumor growth trend data.
9. The intelligent data analysis system before breast tumor resection according to claim 1, characterized in that: The prognostic breast tumor resection risk prediction module includes the following functions: Acquiring historical breast tumor ablation data from a medical database, wherein the historical breast tumor ablation data includes historical breast tumor ablation region data and historical ablation risk label data; Based on the preset convolutional neural network algorithm and historical breast tumor resection data, a relationship prediction model between breast tumor resection area and resection risk is established to construct a breast tumor resection risk prediction model; Performing prognostic breast tumor target area image analysis on breast tumor target area image data according to breast tumor growth trend data to generate prognostic breast tumor target area image data; The prognostic breast tumor target area imaging data is transmitted to the breast tumor cutting risk prediction model to perform prognostic breast tumor cutting risk prediction and generate prognostic breast tumor cutting risk prediction data.
10. The intelligent data analysis system before breast tumor resection according to claim 9, characterized in that: The establishment of a relationship prediction model between breast tumor cutting area and cutting risk based on a preset convolutional neural network algorithm and historical breast tumor cutting data includes: Using historical breast tumor cutting area data as input data and historical cutting risk label data as output data to design a breast tumor cutting risk training set; Based on a preset convolutional neural network algorithm, a mapping relationship between breast tumor resection areas and resection risks is established to construct a breast tumor resection risk prediction model architecture. The breast tumor cutting risk training set is transferred to the breast tumor cutting risk prediction model architecture for model training to obtain a breast tumor cutting risk prediction model.
Citation Information
Patent Citations
Tumor prediction method based on image feature data
CN108847286A
Whole-course intelligent management system, method and equipment for breast tumor patient and medium
CN115458160A
Intelligent tumor disease course monitoring system
CN118522471A
