An artificial intelligence method and system for endometrial cancer screening

Abstract

The application discloses an artificial intelligence endometrial carcinoma screening method and system, and relates to the technical field of endometrial carcinoma screening.The method comprises the following steps: determining risk-related indexes that need to be monitored according to the basic health information of a target screening population; extracting historical risk index distribution characteristics corresponding to different pathological stages from a historical endometrial carcinoma screening database, and extracting risk-related factors between the risk-related indexes and the incidence probability of endometrial carcinoma according to the historical risk index distribution characteristics; extracting target correlation factors from the risk-related factors after processing the current risk index distribution characteristics of the target screening population; obtaining a to-be-analyzed data set by counting original detection data corresponding to the risk-related indexes, and calculating first risk values and second risk values of the risk-related indexes in the to-be-analyzed data set according to the target correlation factors, so that the prevention and treatment effect of endometrial carcinoma is improved.

Description

Technical Field

[0001] This invention relates to the field of endometrial cancer screening technology, specifically to an artificial intelligence-based endometrial cancer screening method and system. Background Technology

[0002] Endometrial cancer is one of the most common malignant tumors of the female reproductive system, posing a significant threat to women's health. It not only severely impacts patients' quality of life but also seriously endangers their lives. Traditional endometrial cancer screening often relies on single detection methods, such as gynecological examinations or endometrial biopsies.

[0003] However, these traditional methods have many limitations. Firstly, some examinations are invasive, such as endometrial biopsies, which require obtaining tissue samples from the uterus using instruments. This can cause discomfort and even pain for the examinee, leading many to resist screening due to fear of such discomfort, resulting in low compliance and delaying timely screening for many potential patients, thus hindering disease detection. Secondly, traditional screening methods often rely on doctors' experience, making them highly subjective. Differences in experience and knowledge among doctors can lead to different interpretations of the same examination results, reducing the effectiveness of prevention and treatment for patients. Summary of the Invention

[0004] To address the technical problem that existing technologies reduce the effectiveness of patient prevention and treatment, this invention provides an artificial intelligence-based method and system for endometrial cancer screening. The technical solution is as follows:

[0005] On the one hand, an artificial intelligence-based method for endometrial cancer screening is provided, which includes the following steps:

[0006] Based on the basic health information of the target screening population, identify the risk-related indicators that need to be monitored.

[0007] The distribution characteristics of historical risk indicators corresponding to different pathological stages were extracted from the historical endometrial cancer screening database. Based on the distribution characteristics of historical risk indicators, risk association factors between risk-related indicators and the incidence probability of endometrial cancer were extracted.

[0008] After processing the current risk indicator distribution characteristics of the target screening population, the target correlation factor is extracted from the risk correlation factor;

[0009] The original detection data corresponding to the statistical risk correlation indicators are used to obtain the data set to be analyzed. The first risk value and the second risk value corresponding to the risk correlation indicators in the data set to be analyzed are calculated according to the target correlation factor.

[0010] The current risk prediction value is obtained by predicting the risk of endometrial cancer in the target screening population based on the first risk value and the second risk value.

[0011] Output high-risk screening notification information based on the current risk prediction value.

[0012] Optionally, the basic health information of the target screening population includes the age information and menstrual information of the target screening population.

[0013] Optionally, risk-related indicators that require key monitoring can be determined based on the basic health information of the target screening population, specifically including the following steps:

[0014] The age information of the target screening population is marked as the primary core indicator;

[0015] Menstrual information of the target screening population was marked as the second core indicator;

[0016] Obtain hormone level test reports and gynecological ultrasound examination data from the target screening population;

[0017] Based on hormone level test reports, gynecological ultrasound examination data, and the first and second core indicators, the risk-related indicators of the target screening population were supplemented and expanded to obtain extended risk indicators.

[0018] The first core indicator, the second core indicator, and the extended risk indicator are combined to form the risk-related indicator.

[0019] Optionally, risk association factors between risk-related indicators and the incidence rate of endometrial cancer can be extracted based on the distribution characteristics of historical risk indicators. This includes the following steps:

[0020] Historical incidence data of risk-related indicators were extracted from the historical endometrial cancer screening database based on the distribution characteristics of historical risk indicators.

[0021] Based on the range of indicator values ​​in the dataset to be analyzed, the risk difference is calculated by performing a difference calculation on the correspondence between risk-related indicators and the probability of disease in historical disease-related data.

[0022] The risk correlation factor is obtained by processing the risk difference and the indicator values ​​of the dataset to be analyzed.

[0023] Optionally, the risk difference and the indicator values ​​of the dataset to be analyzed are processed to obtain the risk correlation factor, specifically including the following steps:

[0024] The risk ratio is calculated by proportionally dividing the risk difference by the range of indicator values ​​in the dataset to be analyzed.

[0025] The risk correlation factor is obtained by weighting all risk ratio values.

[0026] Optionally, after processing the current risk indicator distribution characteristics of the target screening population, the target association factor can be extracted from the risk association factors, specifically including the following steps:

[0027] Obtain the current risk indicator distribution characteristics of the target screening population;

[0028] The matching value is obtained by matching the current risk indicator distribution characteristics with the historical risk indicator distribution characteristics;

[0029] The target correlation factor is extracted from the risk correlation factor based on the matching value.

[0030] Optionally, the first risk value and the second risk value corresponding to the risk correlation indicators in the dataset to be analyzed are calculated based on the target correlation factor, specifically including the following steps:

[0031] The values ​​of risk-related indicators in the dataset to be analyzed are standardized to obtain standardized indicator data.

[0032] The first risk value is obtained by multiplying the standardized indicator data with the target correlation factor corresponding to the first core indicator.

[0033] The second risk value is obtained by multiplying the standardized indicator data and the target correlation factors corresponding to the second core indicator.

[0034] Optionally, a comprehensive risk assessment model is constructed based on the first and second risk values. This model is then used to predict the current risk prediction value for endometrial cancer incidence in the target screening population. The specific steps include:

[0035] The first weight value is obtained by calculating the risk weight coefficient of the first core indicator;

[0036] The first weighted risk value is obtained by multiplying the first risk value and the first weight value.

[0037] The second risk value and the second weight value corresponding to the second core indicator are multiplied together to obtain the second weighted risk value.

[0038] The combined risk value is obtained by summing the first weighted risk value and the second weighted risk value.

[0039] Input the comprehensive risk value into the comprehensive risk assessment model to obtain the current risk prediction value.

[0040] On the other hand, an artificial intelligence-based endometrial cancer screening system is provided, which is applied to an artificial intelligence-based endometrial cancer screening method. The system includes:

[0041] Monitoring module: Determines key risk-related indicators to be monitored based on the basic health information of the target screening population;

[0042] Extraction module: Extracts the distribution characteristics of historical risk indicators corresponding to different pathological stages from the historical endometrial cancer screening database, and extracts risk association factors between risk-related indicators and the incidence probability of endometrial cancer based on the distribution characteristics of historical risk indicators;

[0043] Processing module: After processing the current risk indicator distribution characteristics of the target screening population, extract the target correlation factor from the risk correlation factor;

[0044] Calculation module: Obtain the set of data to be analyzed from the original detection data corresponding to the statistical risk correlation indicators, and calculate the first risk value and the second risk value corresponding to the risk correlation indicators in the set of data to be analyzed based on the target correlation factor;

[0045] Prediction module: Based on the first risk value and the second risk value, the current risk prediction value is obtained by predicting the risk value of endometrial cancer in the target screening population;

[0046] Input module: Outputs high-risk screening notification information based on the current risk prediction value.

[0047] On the other hand, an electronic device is provided, comprising: a processor memory storing computer-readable instructions, wherein when the computer-readable instructions are executed by the processor, an artificial intelligence-based endometrial cancer screening method is implemented.

[0048] On the other hand, a computer-readable storage medium is provided, wherein at least one instruction is stored therein, the at least one instruction being loaded and executed by a processor to implement any one of the methods in an artificial intelligence-based endometrial cancer screening method.

[0049] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0050] This invention determines key risk-related indicators requiring monitoring based on the basic health information of the target screening population. It focuses on critical factors associated with endometrial cancer incidence, avoiding unnecessary indicator testing and thus saving screening costs and time. Historical risk indicator distribution characteristics corresponding to different pathological stages are extracted from a historical endometrial cancer screening database, and risk-related factors are extracted accordingly. Leveraging extensive historical data makes risk assessment more scientific and reliable, more accurately reflecting the relationship between risk-related indicators and incidence probability. Target-related factors are extracted after processing the current risk indicator distribution characteristics of the target screening population, making the risk assessment more relevant to the current situation of the screening population and improving its targeting. By statistically analyzing the raw test data to obtain the dataset to be analyzed, and combining the target-related factors to calculate the first and second risk values, the risk of disease is quantified from different dimensions, making the risk assessment more detailed. Based on the first and second risk values, the current risk value is predicted, and high-risk screening notification information is output. This enables efficient and accurate prediction of endometrial cancer incidence risk, helping to identify high-risk groups in a timely manner, buying time for early intervention and treatment, and thus improving the prevention and treatment of endometrial cancer. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 This is a flowchart of an artificial intelligence-based endometrial cancer screening method provided in an embodiment of the present invention;

[0053] Figure 2 This is a block diagram of an artificial intelligence-based endometrial cancer screening system provided in an embodiment of the present invention;

[0054] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0055] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0056] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0057] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0058] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0059] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0060] This invention provides an artificial intelligence-based method for endometrial cancer screening, which can be implemented by an electronic device, such as a terminal or a server. Figure 1 The flowchart shown is for an artificial intelligence-based endometrial cancer screening method. The processing flow of this method may include the following steps:

[0061] Based on the basic health information of the target screening population, identify the risk-related indicators that need to be monitored.

[0062] The distribution characteristics of historical risk indicators corresponding to different pathological stages were extracted from the historical endometrial cancer screening database. Based on the distribution characteristics of historical risk indicators, risk association factors between risk-related indicators and the incidence probability of endometrial cancer were extracted.

[0063] After processing the current risk indicator distribution characteristics of the target screening population, the target correlation factor is extracted from the risk correlation factor;

[0064] The original detection data corresponding to the statistical risk correlation indicators are used to obtain the data set to be analyzed. The first risk value and the second risk value corresponding to the risk correlation indicators in the data set to be analyzed are calculated according to the target correlation factor.

[0065] The current risk prediction value is obtained by predicting the risk of endometrial cancer in the target screening population based on the first risk value and the second risk value.

[0066] Output high-risk screening notification information based on the current risk prediction value. If the current risk prediction value is greater than or equal to the preset risk warning threshold, output high-risk screening notification information and extract examination plan from the preset diagnosis and treatment suggestion library based on the high-risk screening notification information.

[0067] The AI-powered endometrial cancer screening process utilizes multi-dimensional indicators, such as age, menstrual history, hormone levels, gynecological ultrasound data, obesity, body mass index, hypertension, diabetes, and polycystic ovary syndrome. Hormone levels are presented as range values. These are combined with a comprehensive risk assessment model to calculate the current risk prediction value. This prediction value is a quantitative assessment of the examinee's risk of developing endometrial cancer. The preset risk warning threshold is a key value determined based on extensive clinical data and medical research. If the current risk prediction value is greater than or equal to this threshold, it means that the examinee's risk of developing endometrial cancer has reached a level requiring close monitoring, at which point a high-risk screening notification mechanism is automatically triggered.

[0068] A high-risk screening notification is a warning signal, such as a pop-up notification on the screening system interface or a message sent to relevant medical staff, informing them that the person being screened has a higher risk of developing endometrial cancer and that further examination is needed to confirm the diagnosis.

[0069] The pre-set treatment suggestion database is a collection of recommendations tailored to different disease risk levels and symptom presentations. When a high-risk screening notification is generated, the database retrieves corresponding examination plans based on the high-risk profile indicated in the notification. These plans, developed by medical professionals, include more detailed endometrial biopsies, enhanced magnetic resonance imaging (MRI), and hysteroscopy to more accurately determine whether the patient has endometrial cancer and the specific nature of the disease, providing a basis for subsequent diagnosis and treatment.

[0070] The basic health information of the target screening population includes their age and menstrual information.

[0071] Basic health information for the target screening population includes their age and menstrual history. Generally, the risk of endometrial cancer changes with age; for example, women in the perimenopausal or postmenopausal period may have a relatively higher incidence rate. The regularity of menstruation and the presence of abnormal vaginal bleeding can provide important information for identifying risk-related indicators that require close monitoring.

[0072] Based on the basic health information of the target screening population, key risk-related indicators that need to be monitored are determined, specifically including the following steps:

[0073] The age information of the target screening population is marked as the primary core indicator;

[0074] Menstrual information of the target screening population was marked as the second core indicator;

[0075] Obtain hormone level test reports and gynecological ultrasound examination data from the target screening population;

[0076] Based on hormone level test reports, gynecological ultrasound examination data, and the first and second core indicators, the risk-related indicators of the target screening population were supplemented and expanded to obtain extended risk indicators.

[0077] The first core indicator, the second core indicator, and the extended risk indicator are combined to form the risk-related indicator.

[0078] This invention first labels the age information of the target screening population as the first core indicator and menstrual information as the second core indicator. This is because age and menstrual status are important influencing factors in the incidence of endometrial cancer; for example, older individuals and those with menstrual irregularities often have a relatively higher risk of developing endometrial cancer. Hormone level test reports and gynecological ultrasound examination data of the target screening population are obtained. For example, the secretion of estrogen and progesterone affects endometrial growth, and hormonal imbalances may increase the risk of disease. Gynecological ultrasound examinations can directly show the thickness and morphology of the endometrium; excessively thick endometrium or abnormal echoes indicate a problem. Based on hormone level test reports, gynecological ultrasound examination data, and the first and second core indicators, the risk-related indicators of the target screening population are supplemented and expanded to obtain extended risk indicators.

[0079] For example, an older woman with menstrual irregularities might have elevated estrogen levels in her hormone tests, while a gynecological ultrasound might reveal endometrial thickening. Combining this information allows for the development of more targeted risk indicators. Finally, the first core indicator, the second core indicator, and the extended risk indicators are combined to form a complete risk-related index. This provides a comprehensive and accurate foundation for subsequent feature extraction from historical databases and risk value calculations, helping to more accurately assess the risk of endometrial cancer.

[0080] Based on the distribution characteristics of historical risk indicators, risk association factors between risk-related indicators and the incidence rate of endometrial cancer are extracted, specifically including the following steps:

[0081] Historical incidence data of risk-related indicators were extracted from the historical endometrial cancer screening database based on the distribution characteristics of historical risk indicators.

[0082] The historical endometrial cancer screening database is an information repository that stores a large amount of data accumulated from previous endometrial cancer screenings, such as the age, menstrual status, hormone level test results, gynecological ultrasound examination data of different examinees, as well as whether they were finally diagnosed with endometrial cancer and their pathological stage.

[0083] The distribution characteristics of historical risk indicators are summarized by conducting in-depth analysis of these historical data. For example, the distribution of the incidence rate of endometrial cancer in different age groups, the distribution of the correlation between menstrual abnormalities and the incidence rate, and the distribution of the risk of disease within specific hormone level ranges.

[0084] When extracting historical incidence data for risk-related indicators, the historical incidence data associated with these risk indicators is precisely selected from the historical endometrial cancer screening database based on the summarized distribution characteristics of these risk indicators. For example, if the distribution characteristics of historical risk indicators show that individuals aged 50-60 years with abnormal vaginal bleeding have a higher risk of developing the disease, then relevant data on subjects in this age group with abnormal vaginal bleeding are extracted from the database, including whether they were subsequently diagnosed and their pathological condition at the time of diagnosis. This extracted data constitutes the historical incidence data for risk-related indicators, providing important reference for assessing the current subject's risk of developing the disease.

[0085] Based on the range of indicator values ​​in the dataset to be analyzed, the risk difference is calculated by performing a difference calculation on the correspondence between risk-related indicators and the probability of disease in historical disease-related data.

[0086] After obtaining historical incidence correlation data, and combining it with the indicator value range of the dataset to be analyzed, the risk difference is obtained by performing difference calculation on these historical correspondences.

[0087] First, the dataset to be analyzed has specific numerical ranges for indicators, such as age concentrated between 45 and 55 years old and endometrial thickness between 8 and 15 mm. In contrast, the risk-related indicators in historical disease-related data have a wider range of values ​​and correspond to different probabilities of disease incidence.

[0088] Then, to better adapt historical data to the data to be analyzed, it is necessary to calculate the difference. For example, suppose the historical incidence rate at age 50 was 10%, while the age range in the data to be analyzed is 45-55 years old. The difference between the numerical values ​​of the age range to be analyzed and the historical age of 50, and the corresponding difference in incidence rates, is calculated. This difference is used to obtain the risk difference. The risk difference reflects the degree to which the difference between historical incidence data and the current data to be analyzed in the numerical range of the indicator affects the incidence rate. This provides a basis for subsequent steps such as calculating risk proportions and generating risk association factors, making the risk assessment more closely aligned with the actual situation of the current data to be analyzed.

[0089] The risk correlation factor is obtained by processing the risk difference and the indicator values ​​of the dataset to be analyzed.

[0090] This invention first extracts historical incidence correlation data of risk-related indicators from a historical endometrial cancer screening database based on the distribution characteristics of historical risk indicators. For example, the historical database stores a large number of risk-related indicators for different populations based on age and hormone levels, as well as information on whether these populations ultimately developed endometrial cancer and their incidence probability. This constitutes historical incidence correlation data. Based on the numerical range of the indicators in the dataset to be analyzed, the correspondence between risk-related indicators and incidence probabilities in the historical incidence correlation data is calculated using difference processing to obtain the risk difference. For example, if the numerical range of age indicators in the dataset to be analyzed is 40-50 years old, the historical correspondence between risk-related indicators and incidence probabilities for this age group is examined, and the difference is calculated with the correspondences of other age groups to obtain the risk difference. Finally, the risk difference and the indicator values ​​in the dataset to be analyzed are processed to obtain the risk correlation factor. The risk correlation factor reflects the degree of correlation between the current data to be analyzed and the historical incidence data, providing an important basis for subsequent steps such as calculating risk values.

[0091] Through the first calculation formula The risk association factor F is calculated, where, It is the risk difference corresponding to the i-th sample. It is the risk indicator value in the data set to be analyzed for the i-th sample. is the weight of the i-th sample, and n is the number of samples.

[0092] The risk correlation factor is obtained by processing the risk difference and the indicator values ​​of the dataset to be analyzed, specifically including the following steps:

[0093] The risk ratio is calculated by proportionally dividing the risk difference by the range of indicator values ​​in the dataset to be analyzed.

[0094] The risk correlation factor is obtained by weighting all risk ratio values.

[0095] This invention first calculates the risk ratio by proportionally calculating the risk difference and the numerical range of the indicators in the dataset to be analyzed, thus obtaining a risk ratio value. Assuming the risk difference is 5, and for example, the range of endometrial thickness in the dataset to be analyzed is 10-20, the corresponding risk ratio value is calculated using the appropriate ratio. Next, all obtained risk ratio values ​​are weighted and averaged to obtain the final risk association factor. The weighted averaging takes into account the different levels of importance of different risk ratio values ​​in assessing the risk of endometrial cancer, assigning them different weights before averaging to obtain the risk association factor. The risk association factor reflects the relationship between risk-related indicators and the incidence of endometrial cancer, providing crucial evidence for subsequent risk assessment steps.

[0096] After processing the current risk indicator distribution characteristics of the target screening population, the target association factor is extracted from the risk association factors. The specific steps include:

[0097] Obtain the current risk indicator distribution characteristics of the target screening population;

[0098] The matching value is obtained by matching the current risk indicator distribution characteristics with the historical risk indicator distribution characteristics;

[0099] Let the current risk indicator distribution feature vector be... The historical risk indicator distribution feature vector is Then, through the second calculation formula Calculate the Euclidean distance Through the second calculation formula Calculate the matching value .

[0100] The target correlation factor is extracted from the risk correlation factor based on the matching value.

[0101] This invention first obtains the current risk indicator distribution characteristics of the target screening population. These risk indicators encompass the distribution of factors such as age, hormone levels, and endometrial thickness. For example, the target screening population might be aged 45-55 with fluctuating estrogen levels, constituting the current risk indicator distribution characteristics. The current risk indicator distribution characteristics are then matched with historical risk indicator distribution characteristics in a historical endometrial cancer screening database to obtain a matching value. For instance, historically, there may have been a group of individuals aged 45-58 with similar estrogen levels who were subsequently diagnosed with endometrial cancer at different stages. When comparing the current target population's risk indicator distribution characteristics with these historical characteristics, the greater the similarity, the higher the matching value. Based on the matching value, target association factors are extracted from the risk association factors. A high matching value indicates that the current target population's risk indicator distribution characteristics are more similar to a certain historically highly correlated feature. Therefore, the extracted target association factors reflect the association between the current target population and the probability of endometrial cancer incidence, providing crucial information for subsequent steps such as calculating risk values.

[0102] The calculation of the first and second risk values ​​corresponding to the risk correlation indicators in the dataset to be analyzed, based on the target correlation factors, specifically includes the following steps:

[0103] The values ​​of risk-related indicators in the dataset to be analyzed are standardized to obtain standardized indicator data.

[0104] The first risk value is obtained by multiplying the standardized indicator data with the target correlation factor corresponding to the first core indicator.

[0105] The second risk value is obtained by multiplying the standardized indicator data and the target correlation factors corresponding to the second core indicator.

[0106] First, the values ​​of risk-related indicators in the dataset to be analyzed are standardized to obtain standardized indicator data. Assuming the data includes endometrial thickness of 5-20 mm and hormone levels of 1-10 units, after standardization, these values ​​are transformed into a range similar to 0-1. The standardized indicator data is then multiplied by the target correlation factor corresponding to the first core indicator to obtain the first risk value. If the target correlation factor for the first core indicator of age is 0.6, the standardized endometrial thickness is 0.5, and the standardized hormone level is 0.4, multiplying these standardized values ​​by 0.6 yields the first risk value. Then, the standardized indicator data is multiplied by the target correlation factor corresponding to the second core indicator to obtain the second risk value. For example, the target correlation factor for menstrual information is 0.7, and similarly, the standardized indicator data is multiplied by this to obtain the second risk value. This method obtains risk values ​​associated with different core indicators, laying the foundation for subsequent comprehensive assessment of the risk of endometrial cancer.

[0107] Assuming the target population is women aged 50 years with abnormal vaginal bleeding in the past six months undergoing endometrial cancer screening, the primary core indicator is defined as 50 years old, denoted as A=50, and the secondary core indicator for abnormal menstrual bleeding is denoted as M=1. Hormone level testing showed estrogen (E) at 100 pg / ml (normal range 20-80 pg / ml) and progesterone (P) at 2 ng / ml (normal range 5-20 ng / ml); gynecological ultrasound examination showed an endometrial thickness (T) of 12 mm.

[0108] Data on the correlation between the incidence of endometrial cancer and age in women around 50 years old were extracted from historical endometrial cancer screening databases. The corresponding risk factor F was then calculated. A =0.6.

[0109] The correlation data between the incidence rate of abnormal menstrual bleeding and menstruation in individuals was used to calculate the risk association factor F corresponding to menstruation. M =0.7.

[0110] The woman is currently 50 years old and has menstrual irregularities. Her data shows a high similarity to the high-risk group of women around 50 years old with menstrual irregularities in historical data. Therefore, the target association factor was extracted: age-based target association factor F. A_t =0.6, menstrual target correlation factor F M_t= 0.7.

[0111] The extended indicators are standardized by dividing the portion of each indicator outside the normal range by the maximum value of the normal range, and then summing the results.

[0112] The excess estrogen: (100−80)÷80=0.25.

[0113] The excess of progesterone (below normal, calculated in absolute value): (5−2)÷20=0.15.

[0114] Excess endometrial thickness: (12−5)÷12≈0.583.

[0115] The total standardized index data S = 0.25 + 0.15 + 0.583 = 0.983.

[0116] The first risk value for age is S×F A_t =0.983×0.6≈0.59.

[0117] The second risk value for menstruation is S×F. M_t =0.983×0.7≈0.69.

[0118] Set the first weight value W A =0.4, second weight value W M =0.6.

[0119] The first weighted risk value is 0.59 × 0.4 = 0.236.

[0120] The second weighted risk value is 0.69 × 0.6 = 0.414.

[0121] Overall risk value: 0.236 + 0.414 = 0.65.

[0122] A comprehensive risk assessment model is constructed based on the first and second risk values. This model is then used to predict the current risk prediction value for endometrial cancer in the target screening population. The specific steps include:

[0123] First, we extensively collected clinical data from patients with endometrial cancer and healthy women, including indicators such as age, menstruation, hormones, and ultrasound, as well as pathological diagnostic results. At the same time, we preprocessed the data, filled missing values ​​using methods such as mean and median, identified and processed outliers using box plots and Z-scores, and then standardized or normalized continuous indicators to eliminate dimensional differences.

[0124] Important features related to the onset of the disease were screened using statistical methods and machine learning algorithms. The statistical methods included analysis of variance and chi-square test, while the machine learning algorithm was recursive feature elimination. Principal component analysis was used to reduce dimensionality or new features were constructed by combining clinical knowledge to better reflect the combined effects of risk factors.

[0125] To address the binary classification of endometrial cancer risk, various models were employed, including logistic regression, random forest, support vector machine, and neural network. Preprocessed data was divided into training and testing sets. Taking random forest as an example, the optimal hyperparameters were determined through grid search or random search combined with cross-validation, and the model was trained using the training set.

[0126] The model performance is evaluated on the test set using metrics such as accuracy, precision, recall, F1 score, ROC curve, and AUC. If the model performs poorly on a certain class of samples, the sample weights can be increased or the feature engineering method can be adjusted. Alternatively, multiple models can be fused using voting or stacking methods to improve performance.

[0127] Integrate qualified models into the screening system so that they can automatically receive data and output risk prediction values. At the same time, as new clinical data accumulates, the model is regularly retrained and updated to adapt to changes in disease characteristics and populations, ensuring that the model remains accurate and effective.

[0128] The first weight value is obtained by calculating the risk weight coefficient of the first core indicator;

[0129] The first weighted risk value is obtained by multiplying the first risk value and the first weight value.

[0130] The second risk value and the second weight value corresponding to the second core indicator are multiplied together to obtain the second weighted risk value.

[0131] The combined risk value is obtained by summing the first weighted risk value and the second weighted risk value.

[0132] Input the comprehensive risk value into the comprehensive risk assessment model to obtain the current risk prediction value.

[0133] This invention first calculates the risk weight coefficient of the first core indicator to obtain a first weight value. Different first core indicators have varying degrees of importance in assessing the risk of endometrial cancer, thus requiring weight calculation. For example, for older individuals, the age core indicator has a higher weight in risk assessment, resulting in a first weight value of 0.6. The first risk value and the first weight value are multiplied to obtain a first weighted risk value. Assuming the first risk value is 0.4, the first weighted risk value is 0.4 multiplied by 0.6, resulting in 0.24. Then, the second risk value and the second weight value corresponding to the second core indicator are multiplied to obtain a second weighted risk value. For example, if the second risk value is 0.5 and the second weight value is 0.5, the second weighted risk value is 0.25. The first and second weighted risk values ​​are summed to obtain a comprehensive risk value, resulting in a comprehensive risk value of 0.49.

[0134] First, a large amount of historical data related to endometrial cancer needs to be collected. This data includes primary and secondary risk indicators for different populations, various risk-related indicators such as hormone levels and gynecological ultrasound data, as well as the corresponding final diagnosis of endometrial cancer and pathological stage information. Next, this data undergoes preprocessing, including cleaning out invalid or erroneous data, properly imputing missing values, and encoding categorical data to ensure data quality and usability. Then, machine learning models such as logistic regression, random forest, and support vector machines, or deep learning neural network models are selected. Taking random forest as an example, the preprocessed data is divided into training and test sets. The training set is used to train the random forest model, allowing it to learn the complex relationship between these indicators and the incidence of endometrial cancer. During training, the model continuously adjusts its parameters to minimize prediction error. The test set is used to validate the model's performance, evaluating its accuracy, recall, F1 score, and other metrics to determine if the model can accurately predict the risk of endometrial cancer. If the model performance is unsatisfactory, the algorithm parameters need to be readjusted or a different algorithm needs to be used, and training and validation should be repeated until a model with satisfactory performance is obtained. Through this series of steps, a comprehensive risk assessment model is obtained. This model can accurately predict the risk of endometrial cancer in the target screening population based on input data such as the first weighted risk value and the second weighted risk value.

[0135] Finally, the comprehensive risk value is input into the comprehensive risk assessment model. The comprehensive risk assessment model processes the comprehensive risk value based on a large amount of historical data and a trained algorithm to obtain the current risk prediction value, thereby judging the risk of endometrial cancer in the target screening population.

[0136] Figure 2 This is a block diagram illustrating an artificial intelligence-based endometrial cancer screening system according to an exemplary embodiment. The system is used for an artificial intelligence-based endometrial cancer screening method. (Refer to...) Figure 2 The system includes a monitoring module, an extraction module, a processing module, a calculation module, a prediction module, and an input module. Among them:

[0137] Monitoring module: Determines key risk-related indicators to be monitored based on the basic health information of the target screening population;

[0138] Extraction module: Extracts the distribution characteristics of historical risk indicators corresponding to different pathological stages from the historical endometrial cancer screening database, and extracts risk association factors between risk-related indicators and the incidence probability of endometrial cancer based on the distribution characteristics of historical risk indicators;

[0139] Processing module: After processing the current risk indicator distribution characteristics of the target screening population, extract the target correlation factor from the risk correlation factor;

[0140] Calculation module: Obtain the set of data to be analyzed from the original detection data corresponding to the statistical risk correlation indicators, and calculate the first risk value and the second risk value corresponding to the risk correlation indicators in the set of data to be analyzed based on the target correlation factor;

[0141] Prediction module: Based on the first risk value and the second risk value, the current risk prediction value is obtained by predicting the risk value of endometrial cancer in the target screening population;

[0142] Input module: Outputs high-risk screening notification information based on the current risk prediction value.

[0143] The monitoring module is model IMS-M100. This module has a built-in high-precision data acquisition chip and can connect to hospital information systems and laboratory information management systems. It can automatically retrieve electronic medical records of the target screening population, extract basic health information such as age and menstrual history, and then, combined with a preset risk indicator rule base, quickly determine the risk-related indicators that need to be monitored, such as age range, hormone testing items, and ultrasound examination sites. The preliminary screening results are then transmitted to the extraction module in real time.

[0144] The extraction module is model IMS-E200. It is equipped with a high-performance data processing server (configured with an Intel Xeon E5-2690v4 processor, 64GB of RAM, and 2TB SSD storage), and connects to a historical endometrial cancer screening database (using an Oracle 19c database system) via a dedicated data interface. Using SQL queries and Python data processing scripts, it accurately extracts historical risk indicator distribution characteristic data from the database according to the indicator range determined by the monitoring module, such as incidence probability statistics for different age groups and hormone level combinations. The extracted data is cleaned and format-converted to generate a risk association factor dataset, which is then transmitted to the processing module via an internal high-speed data bus.

[0145] The processing module is an IMS-P300. It utilizes an edge computing server (equipped with an NVIDIA Tesla T4 GPU, 16GB of VRAM) to receive the risk association factor dataset from the extraction module. Then, it employs machine learning algorithms (such as the random forest algorithm, implemented using the Scikit-learn library) to perform similarity matching calculations between the current risk indicator characteristics of the target population (real-time data synchronously collected by the monitoring module, such as current age and recent hormone test values) and historical risk indicator distribution characteristics. The calculation uses a cosine similarity algorithm, sorting the matching results from highest to lowest similarity, extracting the target association factors that best fit the current population, encapsulating them in JSON format, and sending them to the calculation module.

[0146] The computing module is model IMS-C400. It consists of a distributed computing cluster composed of multiple high-performance computing workstations (AMD Ryzen 9 7950X CPU, 32GB memory, equipped with a dedicated mathematical coprocessor). After receiving the target correlation factors from the processing module, the raw test data (such as specific hormone values ​​and endometrial thickness measurements) are first standardized using the Z-score standardization method. Then, combining the target correlation factors corresponding to the first and second core indicators, matrix operations are used to calculate the first and second risk values. The calculation results are temporarily stored in the cluster's shared memory for the prediction module to access.

[0147] The prediction module is model IMS-F500. It utilizes a deep learning server (equipped with two NVIDIA A100 GPUs, 80GB of VRAM per GPU) and deploys a trained comprehensive risk assessment model (a deep neural network model built using the TensorFlow framework). After obtaining the first and second risk values ​​from the calculation module, a weighted risk value is calculated based on preset risk weight coefficients (determined through historical data regression analysis and stored in the model configuration file). The two weighted risk values ​​are then summed to obtain the comprehensive risk value. Finally, the comprehensive risk value is input into the deep neural network model. After forward propagation and activation function calculation, the current risk prediction value is output and sent to the input module via a network interface in numerical form and risk level (low, medium, high).

[0148] The input module is model IMS-I600. It uses an industrial-grade touch screen all-in-one machine (19-inch capacitive screen, 1280×1024 resolution) with a built-in embedded system (based on a customized Linux kernel). After receiving the risk prediction value from the prediction module, it compares it with the preset risk warning threshold (stored in a local configuration database). If the risk prediction value is greater than or equal to the threshold, the high-risk screening notification function is immediately triggered, displaying a prominent high-risk warning message on the touch screen. It also extracts targeted examination plans from the local preset treatment suggestion library (using an SQLite database) based on the risk level and indicator characteristics, such as "hysteroscopy + precise biopsy is recommended." Simultaneously, a notification message is sent to the relevant department doctors through the hospital's internal communication system (such as HL7 protocol).

[0149] The monitoring and extraction modules are connected via TCP / IP over the hospital's intranet. The monitoring module sends determined indicator information to the extraction module using Socket communication. The extraction and processing modules are connected via a high-speed data bus (10Gbps transmission rate), using a custom data transmission protocol to transfer datasets. The processing and calculation modules communicate asynchronously via a distributed message queue (such as RabbitMQ) to ensure reliable data transmission. The calculation and prediction modules interact with each other using shared memory technology (such as Linux's mmap mechanism) for low-latency data exchange. The prediction and input modules establish a long-lived connection via WebSocket protocol to enable real-time data push and display control. The connections between these modules are tight and efficient, forming a complete data flow loop that ensures smooth operation from data acquisition to result output.

[0150] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 3 As shown, electronic devices can be used to achieve Figure 1 The illustrated artificial intelligence-based method for endometrial cancer screening. Optionally, the electronic device may include a first processor 2001.

[0151] Optionally, the electronic device may also include a memory 2002 and a transceiver 2003.

[0152] The first processor 2001, memory 2002, and transceiver 2003 can be connected via a communication bus.

[0153] The following is combined with Figure 3 A detailed introduction to each component of the electronic device:

[0154] The first processor 2001 is the control center of the electronic device and can be a single processor or a collective term for multiple processing elements. For example, the first processor 2001 can be one or more central processing units (CPUs), application-specific integrated circuits (ASICs), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0155] Optionally, the first processor 2001 can perform various functions of the electronic device by running or executing software programs stored in the memory 2002 and calling data stored in the memory 2002.

[0156] In a specific implementation, as one example, the first processor 2001 may include one or more CPUs, for example... Figure 3 CPU0 and CPU1 are shown in the diagram.

[0157] In a specific implementation, as one example, the electronic device 3 may also include multiple processors, for example... Figure 3 The first processor 2001 and the second processor 2004 are shown in the diagram. Each of these processors can be a single-core processor or a multi-core processor. Here, a processor can refer to one or more devices, circuits, and / or processing cores used to process data (such as computer program instructions).

[0158] The memory 2002 is used to store the software program that executes the present invention, and is controlled by the first processor 2001 to execute it. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0159] Optionally, the memory 2002 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory 2002 may be integrated with the first processor 2001 or may exist independently and be coupled to the first processor 2001 through an interface circuit of an electronic device. This embodiment of the invention does not specifically limit this.

[0160] The transceiver 2003 is used to communicate with network devices or with terminal devices.

[0161] Optionally, transceiver 2003 may include a receiver and a transmitter. The receiver is used to implement the receiving function, and the transmitter is used to implement the transmitting function.

[0162] Optionally, the transceiver 2003 can be integrated with the first processor 2001 or exist independently and be coupled to the first processor 2001 through the interface circuit of the electronic device. This embodiment of the invention does not specifically limit this.

[0163] It should be noted that, Figure 3 The structure of the electronic device shown does not constitute a limitation of the present invention. Actual electronic devices may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0164] Furthermore, the technical effects of the electronic device can be referenced from the technical effects of the artificial intelligence endometrial cancer screening method described in the above method embodiments, and will not be repeated here.

[0165] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0166] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0167] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0168] It should be understood that, in various embodiments of the present invention, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0169] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0170] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0171] In the several embodiments provided by this invention, it should be understood that the disclosed devices, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0172] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0173] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0174] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0175] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. An artificial intelligence-based method for endometrial cancer screening, characterized in that, The method includes the following steps: Based on the basic health information of the target screening population, identify the risk-related indicators that need to be monitored. The distribution characteristics of historical risk indicators corresponding to different pathological stages were extracted from the historical endometrial cancer screening database. Based on the distribution characteristics of historical risk indicators, risk association factors between risk-related indicators and the incidence rate of endometrial cancer were extracted. The steps are as follows: Historical incidence data of risk-related indicators were extracted from the historical endometrial cancer screening database based on the distribution characteristics of historical risk indicators. Based on the range of indicator values ​​in the dataset to be analyzed, the risk difference is calculated by performing a difference calculation on the correspondence between risk-related indicators and the probability of disease in historical disease-related data. The risk ratio is calculated by proportionally dividing the risk difference by the range of indicator values ​​in the dataset to be analyzed. The risk correlation factor is obtained by taking a weighted average of all risk ratios. After processing the current risk indicator distribution characteristics of the target screening population, the target association factor is extracted from the risk association factors. The steps are as follows: Obtain the current risk indicator distribution characteristics of the target screening population; The matching value is obtained by matching the current risk indicator distribution characteristics with the historical risk indicator distribution characteristics; The target correlation factor is extracted from the risk correlation factor based on the matching value; The original detection data corresponding to the statistical risk correlation indicators are used to obtain the data set to be analyzed. The first risk value and the second risk value corresponding to the risk correlation indicators in the data set to be analyzed are calculated according to the target correlation factor. The current risk prediction value is obtained by predicting the risk of endometrial cancer in the target screening population based on the first risk value and the second risk value. Output high-risk screening notification information based on the current risk prediction value.

2. The artificial intelligence-based endometrial cancer screening method according to claim 1, characterized in that, The basic health information of the target screening population includes their age and menstrual information.

3. The artificial intelligence-based endometrial cancer screening method according to claim 2, characterized in that, Based on the basic health information of the target screening population, key risk-related indicators that need to be monitored are determined, specifically including the following steps: The age information of the target screening population is marked as the primary core indicator; Menstrual information of the target screening population was marked as the second core indicator; Obtain hormone level test reports and gynecological ultrasound examination data from the target screening population; Based on hormone level test reports, gynecological ultrasound examination data, and the first and second core indicators, the risk-related indicators of the target screening population were supplemented and expanded to obtain extended risk indicators. The first core indicator, the second core indicator, and the extended risk indicator are combined to form the risk-related indicator.

4. The artificial intelligence-based endometrial cancer screening method according to claim 1, characterized in that, The calculation of the first and second risk values ​​corresponding to the risk correlation indicators in the dataset to be analyzed, based on the target correlation factors, specifically includes the following steps: The values ​​of risk-related indicators in the dataset to be analyzed are standardized to obtain standardized indicator data. The first risk value is obtained by multiplying the standardized indicator data with the target correlation factor corresponding to the first core indicator. The second risk value is obtained by multiplying the standardized indicator data and the target correlation factors corresponding to the second core indicator.

5. The artificial intelligence-based endometrial cancer screening method according to claim 4, characterized in that, A comprehensive risk assessment model is constructed based on the first and second risk values. This model is then used to predict the current risk prediction value for endometrial cancer in the target screening population. The specific steps include: The first weight value is obtained by calculating the risk weight coefficient of the first core indicator; The first weighted risk value is obtained by multiplying the first risk value and the first weight value. The second risk value and the second weight value corresponding to the second core indicator are multiplied together to obtain the second weighted risk value. The combined risk value is obtained by summing the first weighted risk value and the second weighted risk value. Input the comprehensive risk value into the comprehensive risk assessment model to obtain the current risk prediction value.

6. An artificial intelligence-based endometrial cancer screening system, applied to the artificial intelligence-based endometrial cancer screening method according to any one of claims 1 to 5, characterized in that, include: Monitoring module: Determines key risk-related indicators to be monitored based on the basic health information of the target screening population; Extraction module: Extracts the distribution characteristics of historical risk indicators corresponding to different pathological stages from the historical endometrial cancer screening database, and extracts risk association factors between risk-related indicators and the incidence probability of endometrial cancer based on the distribution characteristics of historical risk indicators; specifically including: Historical incidence data of risk-related indicators were extracted from the historical endometrial cancer screening database based on the distribution characteristics of historical risk indicators. Based on the range of indicator values ​​in the dataset to be analyzed, the risk difference is calculated by performing a difference calculation on the correspondence between risk-related indicators and the probability of disease in historical disease-related data. The risk ratio is calculated by proportionally dividing the risk difference by the range of indicator values ​​in the dataset to be analyzed. The risk correlation factor is obtained by taking a weighted average of all risk ratios. Processing module: After processing the current risk indicator distribution characteristics of the target screening population, extracts the target correlation factor from the risk correlation factors; specifically including: Obtain the current risk indicator distribution characteristics of the target screening population; The matching value is obtained by matching the current risk indicator distribution characteristics with the historical risk indicator distribution characteristics; The target correlation factor is extracted from the risk correlation factor based on the matching value; Calculation module: Obtain the set of data to be analyzed from the original detection data corresponding to the statistical risk correlation indicators, and calculate the first risk value and the second risk value corresponding to the risk correlation indicators in the set of data to be analyzed based on the target correlation factor; Prediction module: Based on the first risk value and the second risk value, the current risk prediction value is obtained by predicting the risk value of endometrial cancer in the target screening population; Input module: Outputs high-risk screening notification information based on the current risk prediction value.

7. An electronic device, characterized in that, The electronic device includes: processor; A memory storing computer-readable instructions, which, when executed by the processor, implement an artificial intelligence-based endometrial cancer screening method as described in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium contains program code that can be called by a processor to execute an artificial intelligence-based endometrial cancer screening method as described in any one of claims 1 to 5.

Images