Ovarian cancer monitoring system based on iron metabolism related protein

Abstract

The invention provides an ovarian cancer monitoring system based on iron metabolism related protein, comprising: an information monitoring module for monitoring iron metabolism protein index data and ovarian cancer core biomarker index data of a user; the abnormality determination module is used for comparing the iron metabolism protein index data and the ovarian cancer core biomarker index data with historical personal baselines of the current ovarian cancer development stage of the user, determining the variation amplitude, and triggering the ovarian cancer analysis module to perform analysis when the variation amplitude meets preset requirements; and the ovarian cancer analysis module is used for updating the ovarian cancer risk index grade based on the iron metabolism protein index data of the user and the ovarian cancer core biomarker data. According to the invention, the iron metabolism protein index is combined with the ovarian cancer core biomarker, and the biological characteristic of iron-dependent proliferation of the ovarian cancer cells is utilized, so that an index basis which is more suitable for a disease mechanism is provided for monitoring, and the possibility of early abnormality capture is improved.

Description

Technical Field

[0001] This invention belongs to the field of ovarian cancer monitoring technology, specifically relating to an ovarian cancer monitoring system based on iron metabolism-related proteins. Background Technology

[0002] Ovarian cancer, a common and highly dangerous malignant tumor in gynecology, suffers from highly insidious early clinical symptoms due to its biological characteristics. Limited by current medical testing technologies, an early diagnostic system with both ideal sensitivity and specificity has not yet been established, resulting in most patients being diagnosed at an advanced stage, significantly impacting clinical treatment outcomes and patient prognosis. Current clinical practice using ovarian cancer monitoring methods, such as ultrasound imaging and tumor marker testing (e.g., carbohydrate antigen 125), faces significant technical bottlenecks. Specifically, ultrasound examinations are limited by resolution and imaging principles, resulting in insufficient early detection efficacy for small lesions; while serum carbohydrate antigen 125 lacks sufficient sensitivity in early-stage ovarian cancer patients and is prone to false positives in benign gynecological conditions such as pelvic inflammatory disease and endometriosis, failing to meet the practical clinical needs for early ovarian cancer screening. Summary of the Invention

[0003] In view of this, the purpose of the present invention is to provide an ovarian cancer monitoring system based on iron metabolism-related proteins to meet the need to improve the possibility of early abnormality detection.

[0004] To achieve the above objectives, the present invention provides the following technical solution: According to a first aspect, the present invention provides an ovarian cancer monitoring system based on iron metabolism-related proteins, comprising: an information monitoring module for monitoring the user's iron metabolism protein index data and ovarian cancer core biomarker index data; an anomaly determination module for comparing the iron metabolism protein index data and ovarian cancer core biomarker index data with the user's historical personal baseline at the current stage of ovarian cancer development, determining the magnitude of change, and triggering the ovarian cancer analysis module to perform analysis when the magnitude of change meets preset requirements; and an ovarian cancer analysis module for updating the ovarian cancer risk index level based on the user's iron metabolism protein index data and ovarian cancer core biomarker data.

[0005] This embodiment provides an ovarian cancer monitoring system based on iron metabolism-related proteins. It combines iron metabolism protein indicators with core biomarkers of ovarian cancer, leveraging the iron-dependent proliferation of ovarian cancer cells to provide a more accurate indicator basis for monitoring, thus increasing the likelihood of early anomaly detection. Simultaneously, it uses historical individual baseline data as a comparison standard, avoiding the problem of traditional uniform reference values ​​ignoring individual differences. This allows for more accurate identification of abnormal fluctuations in the user's own indicators, reducing misjudgments caused by differences in individual baseline indicators. Finally, a mechanism that triggers analysis based on the magnitude of changes ensures the timeliness of monitoring.

[0006] Other advantages, objectives, and features of the invention will be set forth in the following description and will be apparent to those skilled in the art in some respects, or may be learned by practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0007] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration: Figure 1 This is a flowchart illustrating a specific example of an ovarian cancer monitoring system based on iron metabolism-related proteins, as described in this invention. Detailed Implementation

[0008] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0009] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can also refer to the internal connection of two components; and they can refer to a wireless connection or a wired connection. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0010] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0011] This invention provides an ovarian cancer monitoring system based on iron metabolism-related proteins, such as... Figure 1 As shown, it includes: Information monitoring module 101 is used to monitor the user's iron metabolism protein index data and ovarian cancer core biomarker index data; The anomaly detection module 102 is used to compare the iron metabolism protein index data and the core biomarker index data of ovarian cancer with the user's historical personal baseline of the current stage of ovarian cancer development to determine the magnitude of change. When the magnitude of change meets the preset requirements, the ovarian cancer analysis module is triggered to perform analysis. The ovarian cancer analysis module 103 is used to update the ovarian cancer risk index level based on the user's iron metabolism protein index data and ovarian cancer core biomarker data.

[0012] For example, ovarian cancer cells exhibit iron-dependent proliferation, requiring large amounts of iron for DNA replication and energy metabolism. Therefore, iron metabolism protein indicators are an important indicator in this embodiment for ovarian cancer monitoring. These indicators include membrane iron transporter data, transferrin receptor data, and heavy chain ferritin data. Core biomarker data for ovarian cancer include carbohydrate antigen 125 data and human epididymal protein 4 data. The information monitoring module 101 can be a device connected to a biochemical and immunoassay analyzer, receiving the analyzer's results from the user. Monitoring can be achieved by continuously recording relevant parameters from multiple user tests.

[0013] The anomaly determination module 102 can be embedded hardware, such as any terminal with computing capabilities, or it can exist as a program package. This embodiment does not limit the form of the anomaly determination module, and those skilled in the art can determine it as needed. The current stage of ovarian cancer development can be divided into the screening stage and the suspected case verification stage. The individual baseline performance is different in each stage. When the user's indicator data is significantly different from the historical individual baseline of the current ovarian cancer development stage, that is, the change range meets the preset requirements, the preset requirements can be that the change range of any indicator data exceeds plus or minus 2 standard deviations of the historical individual baseline mean. For example, the historical individual baseline mean of carbohydrate antigen 125 indicator data is 32 U / ml, and the standard deviation is 8 U / ml. Then the preset requirements can be that the change range of carbohydrate antigen 125 indicator data exceeds the range of [16-48] U / ml, which indicates that a stage change may have occurred, and the ovarian cancer analysis module needs to be further triggered for analysis to determine whether to enter the next stage. When the user's indicator data is not significantly different from the historical individual baseline of the current ovarian cancer development stage, that is, the preset requirements are not met, it can be used as monitoring data to update the historical individual baseline. Historical personal baselines can be determined based on the user's previous monitoring data, including historical personal baselines for iron metabolism protein indicators and historical personal baselines for ovarian cancer core biomarker indicators.

[0014] The ovarian cancer analysis module 103, like the anomaly determination module 102, can be embedded hardware or a program package; the form of the ovarian cancer analysis module 103 is not limited here. Based on different stages of ovarian cancer development, and based on the user's iron metabolism protein index data and core biomarker data for ovarian cancer, the ovarian cancer analysis module 103 updates the ovarian cancer risk index level using different methods. Specifically, when the user's current ovarian cancer development stage is the screening stage, the core biomarker data for ovarian cancer includes carbohydrate antigen 125 and human epididymal protein 4 data, and the iron metabolism protein index data includes membrane iron transporter, transferrin receptor, and heavy chain ferritin data. The ovarian cancer analysis module performs the following steps: obtaining the user's menopausal status data; inputting the menopausal status data, iron metabolism protein index data, carbohydrate antigen 125 data, and human epididymal protein 4 data into the first risk assessment model to determine the ovarian cancer risk index level.

[0015] For example, iron accumulation is an important feature of early cancer development. However, women experience a decrease in estrogen after menopause, which leads to reduced iron excretion and thus affects ovarian cancer analysis. Therefore, in this embodiment, menopausal status is included in the risk analysis of ovarian cancer.

[0016] The construction and training process of the first risk assessment model is as follows: First, the data of membrane iron transporter index, transferrin receptor index, heavy chain ferritin index, carbohydrate antigen 125 index, and human epididymal protein 4 index are Z-score standardized. The categorical variable menopausal status is one-hot encoded: premenopausal: [1,0,0], postmenopausal: [0,1,0], perimenopausal: [0,0,1]. The processed feature variables are then constructed into a feature vector with a dimension of 6.

[0017] A logistic regression model was selected as the primary risk assessment model, with ovarian cancer diagnosis as the dependent variable and preprocessed feature vectors as independent variables. The model was trained using 5-fold cross-validation, and the regularization parameters were optimized through grid search to finalize the model parameters. After training, receiver operating characteristic (ROC) curves were used to evaluate model performance. The model's AUC value needed to be ≥0.85, sensitivity ≥80%, and specificity ≥85% before it could be put into use.

[0018] During use, the preprocessed user feature vector, namely the standardized membrane iron transporter index data, transferrin receptor index data, heavy chain ferritin index data, carbohydrate antigen 125 index data, and human epididymal protein 4 index data plus menopausal status data after uniquely heat-encoded, is input into the first risk assessment model after training to calculate the probability of ovarian cancer occurrence; then, the risk index level is divided according to the risk probability, and finally the user's ovarian cancer risk index level in the screening stage is output.

[0019] This invention provides an ovarian cancer monitoring system based on iron metabolism-related proteins. It combines iron metabolism protein indicators with core biomarkers of ovarian cancer, leveraging the iron-dependent proliferation of ovarian cancer cells to provide a more accurate indicator basis for monitoring, thus increasing the likelihood of early anomaly detection. Simultaneously, it uses historical individual baseline data as a comparison standard, avoiding the problem of traditional uniform reference values ​​ignoring individual differences. This allows for more accurate identification of abnormal fluctuations in the user's own indicators, reducing misjudgments caused by differences in individual baseline indicators. Finally, a mechanism that triggers analysis based on the magnitude of changes ensures the timeliness of monitoring.

[0020] As an optional implementation, the ovarian cancer analysis module is also used to: acquire the user's clinical history data; determine the correction coefficient based on the clinical history data; correct the ovarian cancer risk index level based on the correction coefficient to obtain the corrected ovarian cancer risk index level; determine the rate of change of carbohydrate antigen 125 and / or human epididymal protein 4 based on the core biomarker data of ovarian cancer; and perform a second correction on the corrected ovarian cancer risk index level based on the rate of change of carbohydrate antigen 125 and / or human epididymal protein 4, thereby updating the ovarian cancer risk index level.

[0021] For example, in medical history data, benign diseases such as endometriosis and pelvic inflammatory disease can interfere with the carbohydrate antigen 125 index, leading to false positives. Therefore, in the ovarian cancer screening stage, calculating the ovarian cancer risk index level solely based on menopausal status, iron metabolism protein index data, carbohydrate antigen 125, and human epididymal protein 4 index data may lead to biases in risk assessment because it does not take into account individual differences in clinical history and dynamic changes in biomarker indicators.

[0022] In this embodiment, the ovarian cancer analysis module first obtains the user's clinical history data. Based on the degree of influence of different medical histories on the risk of ovarian cancer, a correction coefficient is determined through clinical big data statistical analysis. For example, for users with endometriosis, the correction coefficient is set to 0.8. The correction coefficient is used to correct the initially calculated ovarian cancer risk index level, which can eliminate some of the assessment bias caused by clinical history and obtain a corrected risk index level that is more in line with the user's actual situation.

[0023] Meanwhile, the dynamic changes in core biomarker data for ovarian cancer can more accurately reflect the disease progression trend. For example, a sustained and rapid increase in carbohydrate antigen 125 (CA125) in a short period, even if a single value does not exceed the normal reference range, may indicate an increased risk of ovarian cancer; conversely, if the CA125 value decreases slowly or remains stable, even a slightly higher single value indicates a relatively low risk. Therefore, based on the core biomarker data for ovarian cancer, calculating the rate of change of CA125 and / or human epididymal protein 4 (HEP4) data, and then using this rate of change to further correct the ovarian cancer risk index level, can further improve the accuracy of risk assessment. Specifically, when the rate of change of CA125 is greater than 5 and / or the rate of change of HEP4 is greater than 15, the originally calculated ovarian cancer risk index is multiplied by 1.1, with 1.1 serving as the dynamic risk correction coefficient, thus obtaining the second-corrected ovarian cancer risk index. Based on the second-corrected ovarian cancer risk index, a pre-defined level range is used to obtain the second-corrected ovarian cancer risk index level.

[0024] This invention provides an ovarian cancer monitoring system based on iron metabolism-related proteins. By incorporating clinical history data and determining correction coefficients, it can eliminate the interference of benign diseases such as endometriosis and pelvic inflammatory disease on tumor markers (e.g., false positive elevation of carbohydrate antigen 125), making the risk level more closely reflect the user's actual health status. Furthermore, it performs secondary correction based on the rate of change of core biomarkers, focusing on the dynamic trend of indicators rather than a single value, which can capture early signals of disease progression and avoid delays in intervention due to misjudgment of static indicators. Finally, the two corrections form a dual calibration mechanism of individual medical history + dynamic indicators, reducing the false positive and false negative rates in the screening stage.

[0025] As an optional implementation, the user's current ovarian cancer development stage is the suspected case verification stage; iron metabolism protein index data includes membrane iron transporter index data, transferrin receptor index data, and heavy chain ferritin index data; ovarian cancer core biomarker index data includes carbohydrate antigen 125 index data and human epididymal protein 4 index data; the ovarian cancer analysis module performs the following steps: The system acquires users' menopausal status data; based on various iron metabolism protein indicators, it quantifies the expression intensity ratios of membrane iron transporter, transferrin receptor, and heavy chain ferritin in suspected and target tissues using immunohistochemistry; and inputs the menopausal status data, membrane iron transporter expression intensity ratio, transferrin receptor expression intensity ratio, heavy chain ferritin expression intensity ratio, carbohydrate antigen 125 indicator data, and human epididymal protein 4 indicator data into a pre-trained second risk assessment model to obtain the ovarian cancer risk index level.

[0026] For example, immunohistochemistry is a prior art method and will not be elaborated here. The determination of each expression intensity ratio can be as follows: using Image-ProPlus 6.0 image analysis software, select 5 high-power fields for each slice, and measure the average optical density values ​​of membrane iron transporter, transferrin receptor, and heavy chain ferritin in the suspected cancerous tissue and the target tissue respectively; calculate the expression intensity ratio of each iron metabolism protein: membrane iron transporter expression intensity ratio = average optical density value of membrane iron transporter in suspected cancerous tissue / average optical density value of membrane iron transporter in target tissue; transferrin receptor expression intensity ratio = average optical density value of transferrin receptor in suspected cancerous tissue / average optical density value of transferrin receptor in target tissue; heavy chain ferritin expression intensity ratio = average optical density value of heavy chain ferritin in suspected cancerous tissue / average optical density value of heavy chain ferritin in target tissue. Each ratio is rounded to 3 decimal places.

[0027] The second risk assessment model is constructed and trained as follows: We collected validation data from suspected ovarian cancer cases from multiple centers over the past three years, including a total of 5,000 samples. Among them, 1,500 cases were confirmed to have ovarian cancer by surgical pathology, and 3,500 cases were excluded from the control group by pathological biopsy / follow-up. All samples included complete data on menopausal status, membrane iron transporter expression intensity ratio, transferrin receptor expression intensity ratio, heavy chain ferritin expression intensity ratio, carbohydrate antigen 125, human epididymal protein 4, and pathological diagnosis results.

[0028] The expression intensity ratios of membrane iron transporters, transferrin receptors, heavy chain ferritin, carbohydrate antigen 125, and human epididymal protein 4 were subjected to Min-Max normalization. The menopausal status was numerically mapped to non-menopausal = 0, perimenopausal = 0.5, and menopausal = 1, and finally a feature vector with a dimension of 6 was constructed.

[0029] A random forest model was selected as the secondary risk assessment model. The model parameters were set as follows: 100 decision trees, maximum tree depth 10, minimum number of splits per sample 10, and minimum number of leaf nodes per sample 5. Ovarian cancer was used as the dependent variable, and the preprocessed feature vectors were used as independent variables. The training and test sets were divided in a 7:3 ratio. Parameters were optimized using grid search on the training set, and model performance was validated on the test set. On the test set, the model's AUC value needed to be ≥0.9, sensitivity ≥85%, and specificity ≥90%. The importance of each feature was also calculated to ensure that iron metabolism-related features significantly contributed to the model's predictions. Once the performance requirements were met, the model was determined to be the final usable model.

[0030] When new patient data (feature vector) is input into a trained random forest, this new data sample follows the rules of each decision tree from the root node to the leaf node, obtaining the prediction result of that tree: "It is ovarian cancer" or "It is not ovarian cancer". In this embodiment, the output is the proportion of votes from all decision trees predicting "It is ovarian cancer". For example, if 85 out of 100 trees vote "yes", then the risk probability is 0.85. Finally, based on the risk probability, a risk index level is assigned, and the user's ovarian cancer risk index level in the suspected case verification stage is output.

[0031] This invention provides an ovarian cancer monitoring system based on iron metabolism-related proteins. It uses immunohistochemistry to quantify the expression intensity ratio of iron metabolism proteins between suspected cancerous tissue and target tissue. Compared to serum marker detection, this directly reflects local iron metabolism abnormalities in the lesion, improving the correlation between the marker and the lesion. By adding a lesion-specific dimension through tissue expression intensity ratio, risk assessment shifts from systemic indicators to a combination of local and systemic indicators, making it more suitable for diagnostic scenarios in the validation phase and improving the accuracy of validation.

[0032] The above-described process of using a random forest model to rank ovarian cancer risk indices is a standard random forest process. To further explore the complex biological associations between different biomarkers, the second risk assessment model of this invention further employs the following process to improve the model's discriminative performance: Based on the attention mechanism and the expression intensity ratios of membrane iron transporters, transferrin receptors, and heavy chain ferritin, a weight matrix is ​​generated. The expression intensity ratios of membrane iron transporters, transferrin receptors, and heavy chain ferritin are then weighted according to the weight matrix to obtain corresponding weighted intensity ratios. Each weighted intensity ratio is cross-calculated with menopausal data to obtain an interaction feature set consisting of the interaction features between membrane iron transporter expression intensity ratio and menopausal data, the interaction features between transferrin receptor expression ratio and menopausal data, and the interaction features between heavy chain ferritin expression ratio and menopausal data. Based on carbohydrate antigen 1... By comparing the 25 index data with various weighted intensity ratios, a set of association features was determined, consisting of the association features between the expression intensity ratio of membrane iron transporters and carbohydrate antigen 125 index data, the association features between the expression intensity ratio of transferrin receptors and carbohydrate antigen 125 index data, and the association features between the expression intensity ratio of heavy chain ferritin and carbohydrate antigen 125 index data. The basic features, interaction feature set, and association feature set were input into a random forest model to obtain the ovarian cancer risk index level. The basic features were menopausal status data, membrane iron transporter expression intensity ratio, transferrin receptor expression intensity ratio, heavy chain ferritin expression intensity ratio, carbohydrate antigen 125 index data, and human epididymal protein 4 index data.

[0033] For example, the second risk assessment model is a model that integrates an attention mechanism, a feature engineering construction module, and a random forest algorithm module. The training process for the attention mechanism is as follows: a three-layer neural network is constructed, with 3 nodes in the input layer, 16 nodes in the hidden layer using the ReLU activation function, and 3 nodes in the output layer using the Softmax activation function to ensure that the sum of the output values ​​is 1. Preprocessed training set protein data is used (…). , , Using ovarian cancer disease labels as supervision signals and binary cross-entropy as the loss function, the Adam optimizer (learning rate 0.001) is used for training. During training, 20% of the training set is divided into a validation set, and an early stopping strategy is implemented: training is terminated if the validation set loss does not decrease for 10 consecutive rounds to prevent overfitting. After training, the final attention weight vector is saved. , , ].

[0034] Therefore, in the feature engineering construction module, firstly, a weighted protein expression intensity ratio is calculated for each sample; specifically, the weighted intensity ratio of membrane iron transport proteins is calculated. = × Transferrin receptor weighted intensity ratio = × Heavy chain ferritin weighted strength ratio = × Then, the interaction feature is calculated for each sample, which is the product of each weighted intensity ratio and the menopausal state: = ×M; = ×M; = ×M. This interaction feature set has 3 dimensions. Finally, the association features between each weighted intensity ratio and carbohydrate antigen 125(C) are calculated for each sample. First, the difference association features are calculated: =| -C|, =| -C|, =| -C|. Next, calculate the ratio correlation characteristics: , , ,in, It is a minimum value added to prevent division by zero errors.

[0035] The basic features, interaction feature set, and association feature set of each sample are concatenated to form a complete 15-dimensional feature vector. Based solely on these 15-dimensional fused features from the training set, the mean and standard deviation of each feature dimension are calculated. Then, these parameters are used to perform Z-score normalization on the 15-dimensional fused features of both the training and test sets. Using the normalized 15-dimensional fused features of the training set and their corresponding disease labels, a random forest model is trained. The preferred parameters are: 100 decision trees, maximum depth 15, 4 features considered during splitting, minimum number of samples per leaf node 2, and minimum number of samples per split node 5.

[0036] The above process utilizes attention mechanisms to automatically learn and assign differentiated weights to three iron metabolism proteins, simulating the real-world situation where different proteins contribute differently in vivo. This allows the model to focus on key signals. Furthermore, by multiplying and cross-referencing the weighted protein expression with menopausal status, interactive features are explicitly constructed, thereby capturing complex biological synergistic effects such as a higher risk of abnormal expression of specific proteins in postmenopausal women. Simultaneously, by calculating the difference and ratio association features between proteins and carbohydrate antigen 125, the model is given stronger generalization ability, making it less susceptible to being misled by accidental associations in the training data, thus improving the accuracy of ovarian cancer diagnosis.

[0037] As an optional implementation, an ovarian cancer monitoring system based on iron metabolism-related proteins further includes: a risk index curve generation module for generating an ovarian cancer risk index level curve based on an updated ovarian cancer risk index level; an iron metabolism protein index curve generation module for generating a target iron metabolism protein index curve; a correlation determination module for determining the correlation between the ovarian cancer risk index level curve and the target iron metabolism protein index curve; a risk characteristic parameter calculation module for determining curve characteristic parameters based on the ovarian cancer risk index level curve; and a suggestion generation module for generating medical advice based on the correlation, curve characteristic parameters, and ovarian cancer risk index level.

[0038] For example, with the detection time as the X-axis, the risk index level as the Y-axis, and the original risk probability value as the basis for curve fill color mapping, a line graph is drawn using the plot function of Matplotlib. Each data point is labeled with the detection time and risk probability, and the curve smoothness is set to medium to generate an ovarian cancer risk index level curve.

[0039] Because transferrin receptors have the highest specificity in iron uptake by ovarian cancer cells, they are selected as the core target indicator. Users can also manually select membrane iron transporters or heavy chain ferritin as target indicators via the display module. Historical detection data for the target indicator is retrieved from the information monitoring module, including the detection timestamp, original concentration value, instrument serial number, and reagent batch number. After sorting by detection time, the original concentration values ​​are Z-scored to eliminate the influence of individual baseline differences on the curve trend. Then, a solid line graph is plotted using Matplotlib's `plot` function, with detection time as the X-axis and the standardized indicator concentration value as the Y-axis. Data points are marked with circles, generating a target iron metabolism protein indicator curve.

[0040] The corresponding data points of the risk index curve and the target iron metabolism protein index curve are extracted, ensuring that the detection timestamps of both are completely consistent, thus forming a sample set. The linear correlation between the two is calculated using the Pearson correlation coefficient, outputting the correlation coefficient r and the significance P-value. The correlation is determined based on the correlation coefficient r and the significance P-value, specifically: |r|≥0.7 and P<0.05 indicates a strong correlation; 0.5≤|r|<0.7 and P<0.05 indicates a moderate correlation; 0.3≤|r|<0.5 and P<0.05 indicates a weak correlation; and |r|<0.3 or P≥0.05 indicates no significant correlation. A scatter plot is also plotted to visually display the correlation trend.

[0041] Based on the risk index curve, the curve's characteristic parameters are calculated, including the trend slope k, peak frequency f, and volatility ΔP. Specifically, the trend slope k is calculated as follows: a linear regression is used to fit the curve, and the slope k is calculated using a linear regression tool, which can be a simple linear regression model or a multiple linear regression model. The peak frequency f is calculated by statistically analyzing the risk index levels over the past 6 months. The volatility ΔP is calculated by subtracting the difference between the maximum and minimum risk probabilities in the curve.

[0042] As an optional implementation, it is suggested that the generation module perform the following steps: comparing the correlation, curve feature parameters, and ovarian cancer risk index level with corresponding preset thresholds to determine the parameter combination type that meets the preset thresholds; and based on the parameter combination type, retrieving a preset suggestion template to generate a medical consultation suggestion.

[0043] Specifically, k > 0.02 / week indicates a rapid upward trend, 0.01 / week ≤ k ≤ 0.02 / week indicates a slow upward trend, -0.01 / week ≤ k < 0.01 / week indicates a stable trend, and k < -0.01 / week indicates a downward trend. The number of times the risk index level is ≥ 3 within the past 6 months is counted; f ≥ 3 times indicates a high-frequency peak, 1 ≤ f < 3 times indicates a medium-frequency peak, and f = 0 times indicates no peak. ΔP ≥ 0.3 indicates large fluctuations, 0.1 ≤ ΔP < 0.3 indicates medium fluctuations, and ΔP < 0.1 indicates small fluctuations. These three types of parameters are compiled into a risk characteristic parameter report, including parameter name, calculation result, and trend judgment conclusion. For example, a trend slope k = 0.025 / week indicates a rapid upward trend, a peak frequency f = 2 times indicates a medium-frequency peak, and a fluctuation amplitude ΔP = 0.28 indicates medium fluctuations. The risk index levels are divided into four categories based on the risk probability S: Level 1 is low risk, corresponding to S<0.1; Level 2 is low to medium risk, corresponding to 0.1≤SP<0.3; Level 3 is medium to high risk, corresponding to 0.3≤S<0.5; and Level 4 is high risk, corresponding to P≥0.5.

[0044] Based on the above information, combination types can be formed, such as: Combination Type 1: Strong correlation, rapidly rising trend slope, high peak frequency, large fluctuation range, and risk level 4; Combination Type 2: Moderate correlation, slowly rising trend slope, medium peak frequency, moderate fluctuation range, and risk level 3; Combination Type 3: Weak correlation, stable trend slope, no peak frequency, small fluctuation range, and risk level 2. For Combination Type 1, the preset recommendation template is: It is recommended to visit a gynecologic oncology specialist within 24 hours and undergo a pelvic enhanced MRI examination; For Combination Type 2, the preset recommendation template is: It is recommended to visit a gynecologic outpatient clinic within one week and undergo a pelvic ultrasound examination; For Combination Type 3, it is recommended to undergo a routine gynecological examination within three months, including pelvic ultrasound and carbohydrate antigen 125 testing.

[0045] This invention provides an ovarian cancer monitoring system based on iron metabolism-related proteins. By using risk index curves and iron metabolism protein indicator curves, discrete monitoring data is transformed into visualized trends, facilitating intuitive observation of the correlation between risk changes and iron metabolism indicators. Simultaneously, correlation analysis reveals the intrinsic link between iron metabolism indicators and ovarian cancer risk; for example, the strong correlation between changes in transferrin receptors and risk levels provides direction for subsequent indicator optimization and mechanism research. In this embodiment, risk characteristic parameters (trend slope, peak frequency, etc.) are used to quantify the speed and stability of risk changes, extending risk assessment from level determination to risk progression analysis. This allows for the generation of medical recommendations based on multi-dimensional parameters, avoiding the generality of recommendations caused by a single risk level and improving the adaptability of the recommendations.

[0046] As an optional implementation, an ovarian cancer monitoring system based on iron metabolism-related proteins further includes: a display module for displaying the ovarian cancer risk index level change curve and medical advice; and a communication module for pushing the ovarian cancer risk index level and / or medical advice to a target device. The target device can be a user terminal, such as a mobile phone, tablet, etc.; or it can be a medical terminal device, such as a hospital's HIS system terminal, a doctor's mobile terminal, etc. This embodiment helps ensure that users are promptly aware of their own risks and that medical staff simultaneously obtain monitoring results.

[0047] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention.

Claims

1. An ovarian cancer monitoring system based on iron metabolism-related proteins, characterized in that, include: The information monitoring module is used to monitor users' iron metabolism protein index data and ovarian cancer core biomarker index data; The anomaly detection module is used to compare the iron metabolism protein index data and the core biomarker index data of ovarian cancer with the user's historical personal baseline at the current stage of ovarian cancer development to determine the magnitude of change. When the magnitude of change meets the preset requirements, the ovarian cancer analysis module is triggered to perform analysis. The ovarian cancer analysis module is used to update the ovarian cancer risk index level based on the user's iron metabolism protein index data and core ovarian cancer biomarker data.

2. The ovarian cancer monitoring system based on iron metabolism-related proteins according to claim 1, characterized in that, The user's current ovarian cancer stage is the screening stage. Iron metabolism protein data includes membrane iron transporter data, transferrin receptor data, and heavy chain ferritin data. Core ovarian cancer biomarker data includes carbohydrate antigen 125 data and human epididymal protein 4 data. The ovarian cancer analysis module will perform the following steps: Obtain user menopausal status data; Data on menopausal status, membrane iron transporter index, transferrin receptor index, heavy chain ferritin index, carbohydrate antigen 125 index, and human epididymal protein 4 index were input into the first risk assessment model to determine the ovarian cancer risk index level.

3. The ovarian cancer monitoring system based on iron metabolism-related proteins according to claim 2, characterized in that, The ovarian cancer analysis module is also used for: Obtain the user's clinical medical history data; The correction coefficient was determined based on clinical history data; Based on the correction coefficient, the ovarian cancer risk index level is corrected to obtain the corrected ovarian cancer risk index level. Based on the core biomarker data of ovarian cancer, determine the rate of change of carbohydrate antigen 125 and / or the rate of change of human epididymal protein 4. Based on the rate of change of carbohydrate antigen 125 and / or human epididymal protein 4, the corrected ovarian cancer risk index level is revised a second time to update the ovarian cancer risk index level.

4. The ovarian cancer monitoring system based on iron metabolism-related proteins according to claim 1, characterized in that, The user's current ovarian cancer stage is the suspected case verification stage; iron metabolism protein index data includes membrane iron transporter index data, transferrin receptor index data, and heavy chain ferritin index data; ovarian cancer core biomarker index data includes carbohydrate antigen 125 index data and human epididymal protein 4 index data; the ovarian cancer analysis module performs the following steps: Obtain user menopausal status data; Based on the data of various iron metabolism proteins, the expression intensity ratios of membrane iron transporter, transferrin receptor, and heavy chain ferritin in suspected cancerous tissues and target tissues were quantified by immunohistochemistry. The data on menopausal status, the expression intensity ratio of membrane iron transporters, the expression intensity ratio of transferrin receptors, the expression intensity ratio of heavy chain ferritin, the carbohydrate antigen 125 index data, and the human epididymal protein 4 index data were input into a pre-trained second risk assessment model to obtain the ovarian cancer risk index level.

5. An ovarian cancer monitoring system based on iron metabolism-related proteins according to claim 4, characterized in that, Data on menopausal status, the expression intensity ratios of membrane iron transporters, transferrin receptors, and heavy chain ferritin, as well as data on carbohydrate antigen 125 and human epididymal protein 4, were input into a pre-trained second risk assessment model to obtain an ovarian cancer risk index, including: A weight matrix is ​​generated based on the attention mechanism and the expression intensity ratios of membrane iron transporters, transferrin receptors, and heavy chain ferritin. The expression intensity ratios of membrane iron transporters, transferrin receptors, and heavy chain ferritin were weighted based on a weight matrix to obtain the corresponding weighted intensity ratios. Each weighted intensity ratio was cross-calculated with the menopausal status data to obtain an interaction feature set consisting of the interaction features between the expression intensity ratio of membrane iron transporter protein and the menopausal status data, the interaction features between the expression intensity ratio of transferrin receptor and the menopausal status data, and the interaction features between the expression intensity ratio of heavy chain ferritin and the menopausal status data. Based on the carbohydrate antigen 125 index data and various weighted intensity ratios, a set of association features was determined, consisting of the association features between the expression intensity ratio of membrane iron transporters and carbohydrate antigen 125 index data, the association features between the expression intensity ratio of transferrin receptors and carbohydrate antigen 125 index data, and the association features between the expression intensity ratio of heavy chain ferritin and carbohydrate antigen 125 index data. The basic features, interaction feature set, and association feature set were input into the random forest model to obtain the ovarian cancer risk index level. The basic features were menopausal status data, membrane iron transporter expression intensity ratio, transferrin receptor expression intensity ratio, heavy chain ferritin expression intensity ratio, carbohydrate antigen 125 index data, and human epididymal protein 4 index data.

6. The ovarian cancer monitoring system based on iron metabolism-related proteins according to claim 1, characterized in that, Also includes: The risk index curve generation module is used to generate an ovarian cancer risk index level curve based on the updated ovarian cancer risk index level. The iron metabolism protein index curve generation module is used to generate target iron metabolism protein index curves. The correlation determination module is used to determine the correlation between the ovarian cancer risk index grade curve and the target iron metabolism protein index curve. The risk characteristic parameter calculation module is used to determine the curve characteristic parameters based on the ovarian cancer risk index level curve. The suggestion generation module is used to generate medical advice based on correlation, curve feature parameters, and ovarian cancer risk index levels.

7. An ovarian cancer monitoring system based on iron metabolism-related proteins according to claim 6, characterized in that, Also includes: The display module is used to show the curve of changes in the ovarian cancer risk index level and medical advice. The communication module is used to push ovarian cancer risk index levels and / or medical advice to the target device.

8. An ovarian cancer monitoring system based on iron metabolism-related proteins according to claim 6, characterized in that, It is recommended to generate a module that includes: The correlation, curve feature parameters, and ovarian cancer risk index level are compared with the corresponding preset thresholds to determine the parameter combination type that meets the preset thresholds. Based on the parameter combination type, a preset suggestion template is retrieved to generate a medical consultation suggestion.

Images