Diagnostic markers derived from peripheral blood ctl cells of breast cancer and use thereof

Abstract

This invention belongs to the field of biodetection technology, specifically relating to breast cancer diagnostic biomarkers and their applications. When using the 15 biomarkers provided by this invention for breast cancer diagnosis, it exhibits high accuracy (83.8%), high sensitivity (84.1%), and high specificity (91.9%), enabling efficient early diagnosis of breast cancer (AUC=0.94).

Description

Technical Field

[0001] This invention belongs to the field of biological detection technology, specifically relating to diagnostic markers derived from peripheral blood CTL cells in breast cancer and their applications. Background Technology

[0002] The immune system, as a vital defense system against invading pathogens and the elimination of abnormal cells, plays a crucial role in the development of various chronic diseases, including cancer and aging. Immune surveillance, in particular, can promptly identify and eliminate cancerous cells, and its functional status directly impacts disease progression and the efficacy of subsequent treatments. In breast cancer, a common malignant tumor, immune system imbalance is closely related to tumor development, invasion, metastasis, and prognosis. Therefore, monitoring immune-related indicators for early diagnosis and disease progression assessment of breast cancer has become an important research direction in clinical practice.

[0003] However, current technologies for immune-related testing in breast cancer still have significant shortcomings, making it difficult to meet the actual needs of clinical diagnosis and treatment. Firstly, the mainstream detection method relies on flow cytometry with fluorescently labeled antibodies. While this technology can analyze cell surface markers, the patterns of change are complex, the operation is relatively cumbersome, and it lacks a simple and effective set of standards and methods for evaluating cellular immune function, thus limiting its large-scale clinical application. Secondly, cutting-edge technologies developed in recent years, such as next-generation sequencing (including DNA methylation profiling analysis from immune cells and TCR / BCR diversity detection) and flow cytometry (which can identify 40-60 marker molecules at once), while demonstrating high detection resolution in basic research, remain in the laboratory research stage due to their complex operation, high detection costs, and the need for large amounts of clinical samples for validation. They cannot yet be translated into widely usable clinical detection methods.

[0004] Thirdly, in the early diagnosis and treatment efficacy evaluation of breast cancer, the primary reliance is on imaging techniques such as ultrasound, mammography, and MRI, with a lack of blood biomarkers. Early cancer diagnosis, in particular, has always been a challenge in the field. Furthermore, in immunotherapy, efficacy evaluation mainly depends on imaging diagnosis, which not only has a certain lag but also fails to reflect treatment effects in a timely manner and is prone to false negatives, making it difficult to accurately guide clinicians in adjusting medication regimens. Therefore, developing a simple, low-cost, highly accurate detection technology for early breast cancer diagnosis is of great significance for improving the current state of breast cancer diagnosis and treatment. Summary of the Invention

[0005] Based on this, the purpose of this invention is to provide a diagnostic biomarker derived from peripheral blood CTL cells (cytotoxic T lymphocytes) of breast cancer and its application. The diagnostic biomarker can accurately and quickly achieve the diagnosis of early breast cancer as well as the grading and prognosis of breast cancer.

[0006] To achieve the above objectives, the present invention can adopt the following technical solutions: This invention provides a diagnostic biomarker derived from peripheral blood CTL cells in breast cancer, comprising one or more of the following indicators: CD3 + CD8 + Psi - CD3 + CD8 + CD3 + CD8 + FAS - Psi - CD8 + CD45RA - CCR7 - Psi + CD8 + CD45RA - CCR7 - Psi - CD8 + FAS + Psi + CD8 + CD45RA + CCR7 + Psi + CD8 + CD45RA + CCR7 + Psi - CD8 + TOX + PD1 + CD8 + KLRG1 - PD1 + CD8 + CD45RA + CCR7 - Psi + CD8 + CD45RA + CCR7 - Psi - CD8 + CD45RA - CCR7 + Psi +CD4 + CD45RA - CCR7 + Psi - or CD3 + CD4 + Psi + .

[0007] Preferably, the above applications include: the use of diagnostic marker detection reagents in the preparation of breast cancer diagnostic products; or the use of diagnostic marker detection reagents in the preparation of breast cancer grading products; or the use of diagnostic marker detection reagents in the preparation of breast cancer prognostic products.

[0008] Preferably, in the above applications, the detection reagent for the diagnostic biomarker includes a Psi probe, the structural formula of which is: , where n is any integer from 0 to 7.

[0009] More preferably, in the above application, n is 1.

[0010] Preferably, in the above applications, the detection reagents for diagnostic markers also include one or more of CD3, CD4, CD8, CD45RA, CCR7, KLRG1, PD-1, FAS antibody reagents, phosphate buffered saline solution, or ACK lysis buffer.

[0011] Another aspect of the present invention provides a kit for the diagnosis, grading or prognosis of breast cancer, which includes detection reagents for the diagnostic markers described above.

[0012] Preferably, in the above-mentioned breast cancer diagnosis, grading, or prognosis kit, the detection reagent for the diagnostic biomarker includes a Psi probe, the structural formula of which is: , where n is any integer from 0 to 7.

[0013] More preferably, in the above-mentioned breast cancer diagnosis, grading or prognosis kit, n is 1.

[0014] Preferably, in the above-mentioned breast cancer diagnosis, grading or prognosis kit, the detection reagents for diagnostic markers further include one or more of CD3, CD4, CD8, CD45RA, CCR7, KLRG1, PD-1, FAS antibody reagents, phosphate buffer solution or ACK lysis buffer.

[0015] In another aspect, the present invention provides a breast cancer diagnosis, grading, or prognosis system, comprising: a biomarker acquisition module for acquiring the aforementioned diagnostic biomarkers; a data processing module for processing the acquired diagnostic biomarkers, the processing being based on the XGBoost machine learning method; and an output module for outputting the breast cancer diagnosis, grading, or prognosis.

[0016] The beneficial effects of this invention include: when diagnosing breast cancer based on the 15 biomarkers provided by this invention, it has high accuracy (83.8%), high sensitivity (84.1%) and high specificity (91.9%), and can efficiently achieve early diagnosis of breast cancer (AUC=0.94). Attached Figure Description

[0017] Figure 1 This is a synthesis route diagram for the Psi membrane viscosity probe; Figure 2 This study illustrates the changes in peripheral blood immune cells in mice during tumor progression; (A) is a schematic diagram of the experimental design: female Balb / c mice were inoculated with 5 × 10⁻⁶ cells / mL. 5 4T1 tumor cells, classified into early stage (50–200 mm) based on tumor volume. 3 ), medium term (200–500 mm) 3 ) and late stage (>500mm) 3 (A) Stages; (B) Tumor growth curves of mice at different tumor stages, showing the trend of tumor volume changes over time in the early, middle and late stages; (C) CD3+ in peripheral blood. + CD4 + Flow cytometry analysis of T cell percentage was performed to compare the differences between the normal group and each tumor stage; (D) represents CD3+ in peripheral blood. + CD8 + Flow cytometry analysis of T cell percentage, comparing differences between the normal group and each tumor stage; (E) represents CD8+ in peripheral blood. + Naïve T cells (na Flow cytometry analysis of the proportion of T cells (VE T cells) was performed to compare the differences between the normal group and each tumor stage; (F) represents the proportion of CD8+ in peripheral blood. + Flow cytometry analysis of the proportion of central memory T cells, comparing the differences between the normal group and each tumor stage; (G) represents the CD8+ level in peripheral blood. + Flow cytometry analysis of the proportion of effector memory T cells, comparing the differences between the normal group and each tumor stage; (H) Flow cytometry analysis of the proportion of NK cells in peripheral blood, comparing the differences between the normal group and each tumor stage; *P<0.05, ***P<0.001, ****P<0.0001 (one-way ANOVA, Tukey multiple comparison test; ns: no statistical difference); Figure 3 The relationship between Psi positivity in peripheral blood immune cells of mice and tumor progression; where (A) is a flow cytometry overlay plot showing total CD8+ in peripheral blood of mice in the normal group and at each tumor stage. + Psi +Distribution of mean fluorescence intensity (MFI) in cells; (B) Quantitative statistical analysis, comparing CD8+ in peripheral blood of mice in the normal group and early, middle, and late tumor stages. + Psi + Differences in cell percentage; (C) is a flow cytometry overlay plot showing CD8 levels in peripheral blood of mice in the normal group and at various tumor stages. + T native Psi + Distribution of mean fluorescence intensity (MFI) in cells; (D) represents quantitative statistical analysis, comparing CD8 levels in peripheral blood of mice in the normal group and at each tumor stage. + Naïve T cells Psi + Differences in cell percentage; (E) is a flow cytometry overlay plot showing CD8 levels in peripheral blood of mice in the normal group and at various tumor stages. + Memory T cells (CD8) + T memory Psi + Mean fluorescence intensity (MFI) distribution; (F) represents quantitative statistical analysis, comparing CD8 levels in peripheral blood of mice in the normal group and at each tumor stage. + Memory T cells Psi + Differences in cell percentage. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001; Figure 4 This study investigates the molecular characteristics of peripheral blood CD8+ T cell subsets in breast cancer patients and their relationship with clinical stage. (A) to (D) are scatter plots comparing total CD8+ T cell counts in peripheral blood of a normal control group and breast cancer patients at different stages (Cancer-1, Cancer-2, and Cancer-3), respectively. + T, CD8 + Naïve T cells (CD8) + Tn), CD8 + Central memory T cells (CD8) + Tcm), CD8 + Effector memory T cells (CD8) + Tem) percentage; (E) to (H) are scatter plots and statistical analysis comparisons, respectively representing the total CD8 in peripheral blood of each group. + Psi + Positive cells, CD8 + TnPsi + Positive cells, CD8 + TcmPsi + Positive cells, CD8 + TemPsi +Percentage of positive cells; *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 (one-way ANOVA, Tukey multiple comparison test; ns: no statistical difference); Figure 5 This study investigated the relationship between peripheral blood evaluation indicators and clinical stage in breast cancer patients. (A) Box plot comparing the difference in peripheral blood neutrophil-lymphocyte ratio (NLR) between the normal control group and breast cancer patients at different stages (Cancer-1, Cancer-2, Cancer-3); (B) Box plot comparing the difference in peripheral blood platelet-lymphocyte ratio (PLR) among the groups; (C) Box plot comparing the difference in peripheral blood lymphocyte-monocyte ratio (LMR) among the groups. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001 (One-way ANOVA, Tukey multiple comparison test; ns: no statistical difference). Figure 6 This is a deep learning-based intelligent diagnostic model for breast cancer. (A) is a heatmap of the Pearson correlation coefficients of the top 20 features, showing the linear correlation between each feature (blue indicates positive correlation, red indicates negative correlation); (B) is a bar chart of the top 15 features based on importance ranking (Gain), selecting the key indicators with the highest contribution in the model; (C) is a visualization of the confusion matrix of the three-class model (normal group, stage I, stage II / III breast cancer), comparing the predicted and actual class distributions; (D) is a bar chart of the performance metrics of the three-class model, showing accuracy, precision, sensitivity, specificity, and F1 score; (E) is a visualization of the confusion matrix of the two-class model (normal group, breast cancer group), comparing the predicted and actual class distributions; (F) is a bar chart of the performance metrics of the two-class model, showing accuracy, precision, sensitivity, specificity, and F1 score; (G) is the ROC curve of the two-class model, with an area under the curve (AUC) of 0.9467, evaluating the diagnostic efficacy of the model. Detailed Implementation

[0018] The embodiments described are provided to better illustrate the present invention, but are not intended to limit the scope of the invention to the embodiments described. Therefore, non-essential improvements and adjustments made to the embodiments by those skilled in the art based on the above description are still within the scope of protection of the present invention.

[0019] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. Singular expressions include plural expressions unless they have a distinct meaning in the context. As used herein, it should be understood that terms such as “comprising,” “having,” “including,” are intended to indicate the presence of features, numbers, operations, components, parts, elements, materials, or combinations thereof. The terminology of the invention is disclosed in the specification and is not intended to exclude the possibility that one or more other features, numbers, operations, components, parts, elements, materials, or combinations thereof may be present or added. As used herein, “ / ” may be interpreted as “and” or “or,” depending on the context.

[0020] In a first aspect, embodiments of the present invention provide a diagnostic biomarker derived from peripheral blood CTL cells in breast cancer, the diagnostic biomarker comprising one or more combinations of the following indicators: CD3 + CD8 + Psi - CD3 + CD8 + CD3 + CD8 + FAS - Psi - CD8 + CD45RA - CCR7 - Psi + CD8 + CD45RA - CCR7 - Psi - CD8 + FAS + Psi + CD8 + CD45RA + CCR7 + Psi + CD8 + CD45RA + CCR7 + Psi - CD8 + TOX + PD1 + CD8 + KLRG1 - PD1 + CD8 + CD45RA + CCR7 - Psi + CD8 + CD45RA + CCR7 -Psi - CD8 + CD45RA - CCR7 + Psi + CD4 + CD45RA - CCR7 + Psi - or CD3 + CD4 + Psi + .

[0021] It should be noted that in this invention, the + (positive) and - (negative) in the above-mentioned diagnostic markers are the standardized judgment results of the cell marker expression status / probe binding status by flow cytometry. The core of these results reflects whether the cell being tested expresses a certain type of surface molecule and whether it binds to the Psi membrane viscosity probe. These are the key basis for distinguishing immune cell subtypes and determining the functional status of cells. The specific meanings are explained in two categories: (1) For Psi, + / - indicates the level of cell membrane viscosity after the cell binds to the membrane viscosity probe. It also indicates whether the cell can activate the fluorescence response of the viscosity probe and produce an enhancement effect, reflecting the functional activation status of immune cells. Among them, Psi+ represents that the cell membrane of the target cell is bound to the membrane viscosity probe Psi, and the fluorescence signal intensity of the probe reaches the positive judgment threshold of the flow cytometer, indicating that the cell is in an activated state (the cell membrane viscosity changes, which is the core feature of abnormal T immune cell function in the early stage of tumor); Psi- represents that the target cell is bound to the Psi probe, but the fluorescence signal intensity of the probe is lower than the positive threshold, indicating that the cell does not show obvious activation characteristics and is mostly in a resting / unactivated state; (2) + / - for cell surface molecules (CD3 / CD8 / CD45RA / CCR7 / TOX / PD1 / KLRG1 / CD4 / FAS, etc.): indicates whether the cell expresses the surface marker, where +: represents that the target cell expresses the molecule on its surface, and the fluorescence signal reaches the positive threshold after the molecule is bound by the corresponding fluorescently labeled antibody, which is the core marker for determining cell subtype (such as CD3+ representing that the cell is a T cell, CD8+ representing that it is a CD8). + T cells); - indicates that the target cell does not express / expresses low levels of this molecule on its surface, and the fluorescence signal after antibody binding is below the positive threshold, used to distinguish different cell subpopulations (e.g., CD45RA-CCR7+ is CD8). + Characteristic phenotypes of central memory T cells. For example, CD3+CD8+Psi+ indicates that the cell expresses CD3 and CD8 (defined as CD8). + T cells), and bound to the Psi probe (identified as activated CD8). + (T cells), and so on.

[0022] It should also be noted that CTL cells, or cytotoxic T lymphocytes, are the core effector cells of the body's anti-tumor immunity. Their phenotypic classification and functional status are closely related to the occurrence, development, progression, and immune response level of breast cancer. The diagnostic biomarkers screened in this invention are all specific phenotypic indicators of peripheral blood CTL cells and their subsets in breast cancer, covering cell surface differentiation antigens, functional molecules, immune checkpoint molecules, and Psi marker-related classifications. Large-scale clinical validation has shown that the expression levels of these indicators differ significantly between peripheral blood of breast cancer patients and healthy individuals, and are highly correlated with breast cancer lesion size, lymph node metastasis, clinical stage, and prognosis. A single indicator can achieve preliminary screening, while combined detection of multiple indicators can significantly improve diagnostic accuracy. This can distinguish breast cancer from normal individuals and provide objective evidence for disease grading and prognostic assessment, filling the gap in precise diagnostic biomarkers derived from peripheral blood CTL cells in breast cancer and overcoming the limitations of poor specificity and low sensitivity of traditional tumor markers.

[0023] In some specific examples, the above applications include: the use of diagnostic marker detection reagents in the preparation of breast cancer diagnostic products; or the use of diagnostic marker detection reagents in the preparation of breast cancer grading products; or the use of diagnostic marker detection reagents in the preparation of breast cancer prognostic products.

[0024] It should be noted that the diagnostic biomarker detection reagent of the present invention can specifically bind to the target antigens of CTL cells in peripheral blood, accurately quantify the expression ratio of each biomarker and the proportion of cell subsets, and realize multidimensional clinical applications of breast cancer based on the detection results: when used in diagnostic products, it can realize early non-invasive screening of breast cancer and distinguish between healthy people and early breast cancer patients; when used in grading products, it can accurately determine the clinical stage and pathological grade of breast cancer based on the expression level of each biomarker, reflecting the severity of the disease; when used in prognostic products, it can predict the patient's treatment response, disease-free survival, and overall survival, assess the risk of recurrence, and provide scientific support for the clinical development of individualized treatment plans and follow-up plans. It has comprehensive application scenarios and outstanding clinical practical value.

[0025] In some specific examples, in the above applications, the detection reagents for diagnostic biomarkers include Psi probes, the structural formula of which is: , where n is any integer from 0 to 7.

[0026] It should be noted that the Psi probe is the core functional reagent for the diagnostic biomarker detection of this invention. It can specifically target and bind to the CTL cell membrane, achieving precise labeling of the functional state of CTL cells. It is a key reagent for distinguishing between Psi+ and Psi- cell subpopulations. In the probe's structural formula, n is an integer from 0 to 7, which can regulate the probe's spatial structure, binding affinity, and cell penetration, adapting to the labeling needs of CTL cells with different phenotypes. The probe has the characteristics of strong fluorescence signal, low background noise, no cytotoxicity, and excellent binding ability. It can be adapted to flow cytometry detection platforms and used in conjunction with matching antibody reagents to achieve multicolor fluorescence simultaneous detection, accurately distinguishing various CTL cell subpopulations and ensuring the sensitivity and accuracy of biomarker detection.

[0027] In some specific examples, n is 1 in the above applications.

[0028] It should be noted that when n=1 in the Psi probe structure, it represents the optimal probe structure optimized through experiments. At this point, the probe achieves the best balance between cell penetration efficiency and fluorescence signal stability. It can rapidly penetrate the cell membrane of peripheral blood lymphocytes, accurately bind to the target fragment without background signal interference, and generate a stable and high-intensity fluorescence signal, facilitating accurate identification and quantitative analysis by flow cytometry. Clinical sample validation shows that the Psi probe with n=1 exhibits excellent repeatability and a high signal-to-noise ratio. Compared to other values, it demonstrates superior differentiation between Psi+ and Psi- cell subpopulations, improving detection accuracy by more than 25%. Therefore, it is the optimal Psi probe for detecting the diagnostic biomarkers of this invention.

[0029] In some specific examples, the diagnostic marker detection reagents in the above applications also include one or more of the following: CD3, CD4, CD8, CD45RA, CCR7 antibody reagents, phosphate buffered saline solution, or ACK lysis buffer.

[0030] It should be noted that the above-mentioned auxiliary reagents are all conventional and essential reagents adapted to the detection of diagnostic biomarkers in this invention, ensuring a smooth detection process and accurate results: CD3, CD4, CD8, CD45RA, CCR7, KLRG1, PD-1, and FAS antibody reagents are fluorescently labeled monoclonal antibodies that can specifically bind to the corresponding differentiation antigens on the surface of CTL cells, achieving accurate subtyping of cell populations. When used in conjunction with the Psi probe, they can complete the simultaneous detection of multiple indicators; phosphate-buffered saline (PBS) solution is used for cell washing and resuspending, maintaining the physiological activity and morphology of lymphocytes and avoiding the influence of apoptosis on the detection results; ACK lysis buffer is used to lyse erythrocytes in peripheral blood, remove erythrocyte interference, enrich lymphocytes, and improve the purity of target cell detection. Various reagents can be used alone or in combination, working synergistically with the Psi probe to construct a complete detection system, ensuring the smooth implementation of biomarker detection.

[0031] Secondly, embodiments of the present invention provide a breast cancer diagnosis, grading, or prognosis kit, which includes detection reagents for the diagnostic biomarkers described above.

[0032] It should be noted that this kit uses the detection reagent for peripheral blood CTL cell-derived diagnostic markers for breast cancer of this invention as its core component, integrating a complete set of functional reagents required for detection. It enables one-stop detection of peripheral blood samples without the need for additional core reagents. It is easy to operate, highly versatile, and compatible with flow cytometry platforms in hospital laboratories, third-party testing institutions, and research institutes at all levels. The kit can be flexibly configured with single-indicator or multi-indicator combined detection components according to clinical needs. It can be used for early and rapid diagnosis of breast cancer, as well as for disease grading assessment and prognostic risk prediction. The test results are objective and quantifiable, providing a standardized and convenient tool for precise clinical diagnosis and treatment of breast cancer, and solving the problems of insufficient specificity and limited functionality in existing breast cancer detection products.

[0033] In some specific examples, the diagnostic marker detection reagents in the above-mentioned breast cancer diagnosis, grading, or prognosis kits include Psi probes, the structural formula of which is: , where n is any integer from 0 to 7.

[0034] It should be noted that the Psi probe included in this kit is the core reagent for Psi-related subpopulation typing of CTL cells. The n value in the structural formula ranges from 0 to 7, meeting the personalized detection needs of different detection scenarios and sample types. The probe undergoes purification, exhibiting high purity and good stability, and can be stored long-term at 2-8℃ without significant inactivation. The probe exhibits no cross-reaction or interference with other reagents in the kit, enabling simultaneous multi-indicator fluorescent labeling. Flow cytometry can rapidly obtain expression data for various diagnostic markers, providing intuitive and easily analyzed data, significantly shortening the detection cycle and improving the efficiency of breast cancer diagnosis, grading, and prognostic assessment.

[0035] In some specific examples, n is 1 in the above-mentioned breast cancer diagnosis, grading, or prognosis kits.

[0036] It should be noted that this kit uses a Psi probe with n=1 as the optimal performance probe, which is suitable for the detection needs of most clinical peripheral blood samples. The probe has high binding efficiency and strong fluorescence signal, which can effectively reduce background interference and reduce false positive and false negative results.

[0037] In some specific examples, the diagnostic marker detection reagents in the above-mentioned breast cancer diagnosis, grading or prognosis kits also include one or more of the following: CD3, CD4, CD8, CD45RA, CCR7, KLRG1, PD-1, FAS antibody reagents, phosphate buffered saline solution or ACK lysis buffer.

[0038] It should be noted that all the auxiliary reagents included with this kit have undergone clinical testing and optimization. The antibody reagent is a fluorescently labeled specific monoclonal antibody with stable fluorescence signal and high affinity, enabling simultaneous multicolor flow cytometry typing. The phosphate buffer solution is sterile and calcium- and magnesium-free, suitable for in vitro lymphocyte processing. The ACK lysis buffer is a highly efficient erythrocyte lysis buffer, ensuring thorough lysis without damaging lymphocytes. The reagent groups in the kit are rationally proportioned and are all ready-to-use or easily diluted, requiring no complex pretreatment. Operators can complete the testing after simple training. The kit also includes standards, quality control materials, and detailed operating instructions, enabling end-to-end quality control of the testing process and ensuring accurate and reliable results.

[0039] Thirdly, embodiments of the present invention provide a breast cancer diagnosis, grading, or prognosis system, comprising: a biomarker acquisition module for acquiring the aforementioned diagnostic biomarkers; a data processing module for processing the acquired diagnostic biomarkers, the processing being based on the XGBoost machine learning method; and an output module for outputting the breast cancer diagnosis, grading, or prognosis.

[0040] It is worth noting that this system integrates flow cytometry technology with the XGBoost machine learning algorithm to achieve intelligent and precise analysis of breast cancer diagnosis, grading, and prognosis. The three modules work collaboratively, completing the entire process automatically without manual intervention, significantly improving diagnostic and treatment efficiency. The biomarker acquisition module can interface with a flow cytometer to automatically collect data such as the expression percentage of various diagnostic biomarkers and the proportion of cell subpopulations in peripheral blood CTL cells, achieving precise data acquisition and transmission. The data processing module incorporates an XGBoost model trained and optimized with massive amounts of clinical data. This model has strong capabilities in processing high-dimensional biological data, high prediction accuracy, and excellent anti-overfitting performance. It can quickly uncover the intrinsic correlation between biomarker data and breast cancer disease, enabling precise determination of disease diagnosis, staging, recurrence risk, and prognosis. The output module can output the analysis results in the form of a visual report, clearly presenting diagnostic conclusions, disease grading, prognostic scores, and clinical recommendations. This provides clinicians with an intuitive and scientific basis for developing individualized treatment plans, promoting the intelligent and precise upgrading of breast cancer diagnosis and treatment.

[0041] To better understand the present invention, specific examples are provided below to further illustrate the content of the present invention, but the content of the present invention is not limited to the examples below.

[0042] In the following examples, the synthetic route for the membrane viscosity probe Psi is as follows: Figure 1As shown, the specific preparation method is as follows: (1) Synthesis of intermediate: 1 mmol of 2-(4-pyridyl)acetonitrile (the compound shown in Formula II) and 2 mmol of 4-dimethylaminobenzaldehyde (the compound shown in Formula I (n=1)) are dissolved in acetonitrile and heated under nitrogen protection. The reaction temperature is 90℃ and the reaction time is 6h. After the reaction is completed, the reaction system is cooled and filtered to collect the yellow solid, which is the intermediate (as shown in Formula III (n=1)). (2) Synthesis of fluorescent probe: 1 mmol of the intermediate synthesized in step (1) above and 1 mmol of 1,3-propanesulfonyl lactone (the compound shown in Formula IV) are dissolved in anhydrous ethanol and heated under nitrogen protection. The reaction temperature is 90℃ and the reaction time is 4h. After the reaction is completed, the solvent is removed by rotary evaporator. The crude product is subjected to silica gel column chromatography to obtain a dark red solid, which is the membrane viscosity probe Psi (as shown in Formula V (n=1)).

[0043] Example 1 (I) Animal tumor model experiments at different disease stages (1) MC38, CT26, and 4T1 cells in logarithmic growth phase were resuspended in PBS and their concentrations were adjusted to 1×10⁻⁶. 6 cells / 100μL (MC38 / CT26), 5×10 5 100 μL of tumor cells / 100 cells (4T1); After disinfection of the right back of mice, tumor cell suspension (200 μL per mouse) was subcutaneously injected using a 1 mL syringe. The control group without tumors was injected with an equal volume of PBS. The mental state of the mice was observed daily after inoculation, and the tumor volume was measured with calipers every 3 days. Changes in mouse weight were also recorded. Figure 2 A represents the experimental procedure and tumor stage: Day 0, 5 × 10⁵ cells were injected. 5 4T1 cells, 15 days, 25 days, and 35 days correspond to the early stage (50–200 mm) respectively. 3 ), medium (200–500 mm) 3 ), later stage (>500 mm) 3 ); Figure 2 B represents the growth curve: the tumor grows slowly in the first two weeks, then proliferates rapidly. The late-stage group (green) consistently has the largest tumor volume, followed by the intermediate-stage group (red), and the early-stage group (blue) has the smallest. At 34 days, the tumor volume in the late-stage group is approximately 720 mm. 3 .

[0044] (2) Based on preliminary experimental verification, tumors were classified into early stage (tumor volume 50 mm) according to their size. 3 -200mm 3 ) / Intermediate stage (tumor volume 200mm) 3 -500mm 3 Late stage (tumor volume > 500mm)3 In mice at early, middle, and late stages of tumor development, 200 μL of peripheral blood was collected using the orbital venous plexus sampling method. Pre-added EDTA-K2 anticoagulant EP tubes were used for collection. Immediately after collection, the blood was gently inverted to mix and prevent clotting. The anticoagulant peripheral blood was transferred to a 15 mL sterile centrifuge tube, 800 μL of ACK erythrocyte lysis buffer was added, and the tubes were incubated at room temperature for 15 min, centrifuged at 1000 rpm for 5 min, and the supernatant was discarded. The cells were washed twice with flow cytometry staining buffer (PBS + 2% FBS, the same below) (centrifuged at 1000 rpm for 5 min each time), and finally resuspended in 100 μL of flow cytometry staining buffer (cell concentration approximately 1 × 10⁻⁶). 6 (cells / 100μL) (3) Add the antibodies (100 μL final volume per tube) to the pretreated cell suspension according to the antibody instructions and mix gently: PE-Cy5-anti-mouseCD45, FITC-anti-mouseCD3, PE-anti-mouseCD4, APC-anti-mouseCD8, PerCP-Cy5.5-anti-mouse NK1.1, PE-Cy7-anti-mouse CD49b, FITC-anti-mouse CD44, PE-anti-mouse CD62L, PE anti-mouse CD19, and membrane viscosity probe Psi; incubate at 4℃ in the dark for 20 min. (4) After incubation, add 1 mL of flow cytometry staining buffer, centrifuge at 1000 rpm for 5 min, and discard the supernatant; repeat washing once, resuspend in 300 μL of flow cytometry staining buffer, place the flow cytometry tube on the instrument sample loading platform, repeat the detection 3 times for each sample, and save it in FCS format for subsequent analysis. (5) Use FlowJo software to calculate the percentage of different immune cells, the percentage change of Psi+ probe cells and MFI to reflect the expression level of cell surface markers.

[0045] The results showed that as tumors progressed in the 4T1 mouse model, CD4 levels in peripheral blood decreased. + The percentage of T cells gradually decreased. Figure 2 C), and CD8 + The percentage of T cells gradually increased ( Figure 2 D); Further analysis of CD8 + T cell subsets were found in peripheral blood, including CD8. + Naïve T cells (such as CD44) - CD62L + The percentage of phenotypes (from 72.3% to 26.7%) decreased continuously with tumor progression. Figure 2 E), CD8+ Central memory T cells (Tcm, such as CD44) + CD62L + Phenotypic percentage (from 27.7% to 70.1%) Figure 2 F) and CD8 + Effector memory T cells (Tem, such as CD44) + CD62L - The percentage of phenotypes (from 0% to 2.43%) increased significantly with tumor progression. Figure 2 G); The percentage of NK cells in peripheral blood (from 3.69% to 0.16%) showed a gradual decreasing trend with tumor progression. Figure 2 H); As shown in Figure 3, as the tumor progresses from the early to the middle and late stages, the total CD8+ level in peripheral blood increases. + T cells (Figures 3A and 3B) and their subsets; naïve CD8 + T cells (CD8) + T naive (Figures 3C and 3D) and Memory CD8 + T cells (CD8) + T memor The membrane viscosity probe fluorescence intensity (MFI) and the proportion of probe-positive cells in Figures 3E and 3F showed a progressively increasing trend. The proportion of probe-positive cells in each stage of the tumor disease was significantly higher than that in the normal control group (without tumor), and this trend was highly synchronized with the increase in tumor volume and the progression of the disease.

[0046] It is worth emphasizing that the value of membrane viscosity probes in evaluating immune cell function and in early tumor diagnosis is reflected in their compatibility with "direct CD4 counting". + T, CD8 + The essential difference between T (different subtypes), NK cells, etc.—immune cell counts are more of a "description of the outcome changes of the immune system during tumor progression," such as CD4 cells that only appear in mid-to-late stage tumors. + Decreased T cells and reduced NK cells are not significant in the very early stages of tumors, and may even show no statistical difference from healthy individuals, making it difficult to meet the core requirement of early diagnosis: "early detection." Membrane viscosity probes, on the other hand, detect changes in activated immune cells, particularly CD8. +As a major type of cell-mediated immunity, CTLs (cell-mediated cytotoxicity) can activate specific immune surveillance in both local and systemic environments, even in the very early stages of a small tumor. This is due to changes in tumor cell proliferation and metabolism, alterations in the local microenvironment's metabolites, and the transcriptional activation of certain chemokines and other innate immune signals, or the generation of small amounts of tumor-specific neoantigens. While detection of a large number of effector CTLs in peripheral blood showed no significant change in their overall proportion, nor did it reveal a marked change in the initial resting state of the CD8-T subset, the Psi cell membrane-targeting probe synthesized in this patent is designed to detect cell membrane viscosity. Activation signals from immune cells increase the fluorescence intensity of this probe. It can serve as a functional probe for evaluating immune cells, enabling more sensitive identification of the functional states of different immune cell subsets.

[0047] Flow cytometry analysis of peripheral blood immune cells revealed that CD8+ was present in the aforementioned CTL cells closely related to tumor immune surveillance. + Tnaive, CD8 + Tcm and CD8 + Among the three main types of Tem cells, after loading Psi functional probes, each subpopulation exhibited a certain degree of heterogeneity, with varying proportions of positive and negative cells. This indicates that the probes can further enhance the differentiation ability of each cell subpopulation, especially in the early stages of tumor development, particularly for CD8 cells. + Tnaive cells, compared to normal controls, showed a change in the ratio of positive to negative cell subsets (from 57% / 43% to 82% / 18%). As the tumor progresses, CD8... + Tnaive, gradually moving towards CD8 + TCM and CD8 + Tem differentiation, while CD8 + TCM and CD8 + In Tem, the ratio of Psi positive to negative showed more significant changes and was closely related to tumor volume and disease course, exhibiting high synchronization. This means that membrane viscosity probes can capture early tumor signals where "the counts associated with existing immune cell marker subtypes have not yet fluctuated significantly," addressing the pain point of "late signal appearance" in early diagnosis.

[0048] (II) Clinical Sample Analysis (1) This study strictly followed the established inclusion criteria (age 20-80 years; no history of immunotherapy or vaccination; no history of immunodeficiency diseases; no history of cancer-related surgery; no infectious diseases), systematically collected a cohort of 30 healthy controls (age 20-60 years matched with the patients' ages), and established a cohort of 60 patients with breast cancer diagnosed by pathological biopsy. All patient diagnoses were based on pathological biopsy, the "gold standard," ensuring the reliability and standardization of the research samples. From the TNM staging distribution, the disease severity of the 60 patients showed a clear gradient: 27 patients were in stage I, accounting for 45.0%, the highest proportion among all stages, indicating a large proportion of early-stage breast cancer patients in the cohort; 20 patients were in stage II, accounting for 33.3%, the second highest; and 13 patients were in stage III, accounting for 21.7%, with a relatively low proportion of late-stage patients. This baseline characteristic, with a predominance of early-stage patients and a reasonable distribution of mid-to-late-stage patients, can cover breast cancer patients at different disease progression stages, providing support for the universality of the research results. Meanwhile, clear stage proportion data also provides a clear baseline reference for subsequent comparison of the efficacy differences between the new method and the traditional method, making it easier to more intuitively verify the advantages of the new method in improving diagnostic efficiency and treatment outcomes.

[0049] (2) Take a 1.5 mL sterile centrifuge tube, add 200 μL of EDTA-anticoagulated peripheral blood, and then add 800 μL of ACK red blood cell lysis buffer (the blood: lysis buffer volume ratio is 1:4 to ensure complete lysis of red blood cells). Gently invert to mix, and incubate at room temperature in the dark for 5 minutes (observe the liquid change from red to transparent, indicating that the red blood cells have been completely lysed). Add 1 mL of flow cytometry staining buffer to the centrifuge tube, centrifuge at 350 g for 5 minutes (4℃), and discard the supernatant (to remove lysed red blood cell fragments). Repeat the washing once (centrifuge at 350 g for 5 minutes), and resuspend the precipitate with 100 μL of flow cytometry staining buffer (the precipitate is a white blood cell population, containing immune cells such as T cells). Add immune cell-related flow cytometry antibodies (PE-anti-human CD3 (Biolegend, catalog number: 300308), FITC-anti-human CD4 (Biolegend, catalog number: 317408), APC-anti-human) to the above 100 μL white blood cell suspension according to the antibody instructions. CD4 (Biolegend, part number: 317416), APC-anti-human CD8 (Biolegend, part number: 301014), FITC-anti-human CD45RA (Biolegend, part number: 304106), PE-anti-human CCR7 (Biolegend, part number: 535203), FITC-anti-human CD11b (Biolegend, part number: 101206), PE / Cy7-anti-human CD45 (Biolegend, part number: 368532), FITC-anti-human CD56 (Biolegend, part number: 362545), FITC-anti-human PD-1 (Biolegend, part number: 335709), PE-anti-human PD-1 (Biolegend, part number: 329905), PE-anti-huamn TOX (Invitrogen, Cells were prepared by gently tapping the side of a centrifuge tube with 1µL of a 1mM membrane viscosity probe (Psi, catalog number 50-6502-80), APC-anti-human KLRG1 (Biolegend, catalog number 368605), PE-anti-human FAS (Biolegend, catalog number 305607), and 1µL of a 1mM membrane viscosity probe (Psi). The tube was then incubated at 4°C in the dark for 20 minutes. After incubation, 1mL of flow cytometry staining buffer was added, and the tube was centrifuged at 350g for 5 minutes. The supernatant was discarded. The cells were resuspended in 300μL of flow cytometry staining buffer (cell concentration approximately 1×10⁻⁶). 6 Add cells / mL to flow cytometry tubes, and collect 1×10⁻⁶ cells / mL per sample. 5Each live cell was used to record raw data and save it in FCS format.

[0050] (3) Compare the changes in immune cells and membrane viscosity probe Psi in the tumor patient group and the healthy control group.

[0051] The results are as follows Figure 4 Based on the above flow cytometry detection method, the functional status of T cells in clinical peripheral blood samples can be accurately evaluated. Furthermore, there were significant differences between the tumor patient group and the healthy control group in the composition of peripheral blood immune cells and the positive rate of membrane viscosity probes. Specific results are as follows: Peripheral blood CD3 in cancer patients + The proportion of T cells in the total white blood cell count was 12%-23.50%, lower than that in the healthy control group (17.1%-25.1%); the CD8+ level in the healthy control group was lower. + T cells were 1.08 times higher in the tumor patient group and CD8 in the healthy control group. + Native T cells were 1.01 times higher in the cancer patient group than in the tumor patient group; CD8+ in the cancer patient group + The proportions of Tcm and Tem cells were 1.58 and 1.16 times higher than those in the healthy group, respectively. Compared with the healthy control group (Normal), the total CD8+ in peripheral blood of patients at each stage of tumor progression (Cancer-1, Cancer-2, Cancer-3) was significantly higher. + There was no significant difference in the proportion of T cells. Figure 4 A); and CD8 + T cell subset distribution undergoes significant reprogramming as the disease progresses: initial CD8 + T cells (CD8) + T naive The proportion gradually decreased. Figure 4 B), Central Memory CD8 + T cells (CD8) + Tcm (tumor cytopenia) briefly rises in the mid-stage of cancer (Cancer-2) and then falls back in the late stage (Cancer-3). Figure 4 C) Effect memory CD8 + T cells (CD8) + Tem showed a progressively increasing trend, and was significantly higher in the late stage than in the control group and other tumor stages. Figure 4 D).

[0052] It is noteworthy that, as an early-stage cancer population, stage I breast cancer patients did not show significant differences in peripheral blood lymphocyte markers compared to healthy controls: stage I patients had CD8... + T cell percentage, CD8 +Lymphocyte-related indicators, such as the percentage of native T cells, show a high degree of overlap with the range of values ​​in healthy controls, making it difficult to effectively distinguish between stage I patients and healthy individuals using these traditional immune indicators. However, CD8... + T cells, CD8 + The detection parameters of the Psi membrane viscosity probe in Tnaive cells showed significant differences between stage I patients and healthy controls. Figure 4 E, 4F), among which stage I breast cancer patients group CD8 + The percentage of Tnaive Psi membrane viscosity probe-positive cells was 5.8 times that of the healthy control group, a difference far greater than the changes in lymphocyte-related indicators. This is consistent with tumor-bearing mouse models. In addition, CD8... + T Central Memory、 CD8 + T Effector Memory The positive rates of membrane viscosity probes for cell subsets all showed a progressively increasing trend with the progression of the disease. Figure 4 G, 4H).

[0053] (4) Compare the changes in routine peripheral blood test indicators between cancer patients and healthy individuals. like Figure 5 As shown, complete blood count data were collected from healthy individuals and patients, and the neutrophil-lymphocyte ratio (NLR, Figure 5A), platelet-lymphocyte ratio (PLR, Figure 5B), and lymphocyte-monocyte ratio (LMR, Figure 5C) were calculated and compared. Comparing healthy individuals and patients, it was found that the NLR ratio showed a certain increasing trend compared to the other two values ​​as the tumor progressed, exhibiting the highest relative sensitivity. However, the change was not significant between early-stage breast cancer and healthy individuals.

[0054] Therefore, compared with existing immune markers, such as CD8 + T-cell markers (the fold difference between healthy individuals and stage I patients was only 1.08 times) and CD8 + Compared to naive T cell markers (no significant difference between healthy individuals and stage I patients) and NLR ratios (no significant difference between healthy individuals and stage I patients), only the Psi membrane viscosity probe for different CD8+ T cell subtypes showed high sensitivity in differentiating stage I patients from healthy individuals. It was able to accurately detect CD8+ T cells in the early stages of stage I breast cancer patients, before significant changes in lymphocyte markers occurred. + Abnormal changes at the Tnaive cell functional level provide a more sensitive and reliable detection basis for the early diagnosis of breast cancer, effectively making up for the limitations of traditional immune indicators in early tumor screening, and highlighting the role of membrane viscosity probes in the detection of CD8+. + The key value of T subtype detection in early tumor diagnosis.

[0055] Further stratified analysis by tumor TNM staging showed that the positive rate of membrane viscosity probes in patients with advanced tumors (stages II and III) (54.12%-58.55%) was significantly higher than that in patients with early-stage tumors (stage I, 29.12%). This result suggests that as the tumor progresses, the body continuously mobilizes and activates naive CD8+ CTL cells in response to tumor antigen stimulation (manifested as an increased probe positivity rate). However, activated naive T cells are more likely to differentiate into effector T cells, or gradually show a tendency to be exhausted due to long-term exposure to the tumor microenvironment, ultimately manifesting as the characteristic of "Tnaive cell activation level increasing with tumor progression". By combining the above-mentioned Psi functional probes with their proportion and mutual evolution in CTL cells at different differentiation levels, and comparing them with existing clinical TNM classifications, a precise diagnostic model for breast cancer can be created using deep learning methods.

[0056] Example 2 (1) The number of 24 biomarkers (as shown in Table 1) in peripheral blood clinical samples from tumor patients and healthy controls (a cohort of 60 breast cancer patients diagnosed by pathological biopsy, selected according to the criteria of Example 1) was used as a feature to obtain an initial quantitative data matrix; Table 1 Information on 24 biomarkers

[0057] (2) The extreme gradient boosting (XGBoost) algorithm was selected as the core classifier. This algorithm integrates multiple weak classifiers (decision trees) to gradually reduce the prediction error in an additive model, and introduces a regularization term to control the model complexity, effectively reducing the risk of overfitting; (3) Hyperparameter optimization is performed using a combination of k-fold cross-validation and grid search, specifically including the following steps: (3-1) Randomly divide the training set into 10 mutually exclusive subsets of the same size; (3-2) Each time, select 9 subsets as training subsets and the remaining 1 subset as validation subset; (3-3) Train the model by traversing all possible parameter combinations within the preset hyperparameter space; (3-4) Record the performance metrics of each parameter group on the validation set and select the parameter combination with the best performance. The core hyperparameters for optimization and their value ranges include: learning rate (eta): 0.01-0.3, controlling the weight reduction coefficient of each weak classifier; maximum tree depth (max_depth): 3-10, controlling the maximum depth of a single decision tree; subsample ratio (subsample): 0.6-1.0, controlling the sampling ratio of training instances; feature sampling ratio (colsample_bytree): 0.6-1.0, controlling the sampling ratio of features; minimum child weight (min_child_weight): 1-10, controlling the minimum heuristic weight required for tree splitting; regularization parameter (lambda): 1-10, controlling the weight of the L2 regularization term; number of iterations (n ​​rounds): 100-2000, and using early stopping to prevent overfitting; (4) After training, extract the Gain importance index built into the XGBoost model; Gain represents the average loss function reduction (i.e. average contribution) brought about by a feature as a split point in all trees, which can most accurately reflect the feature prediction ability. Figure 6 A is a heatmap of the Pearson correlation coefficients for the first 20 features, displayed in a downward triangular format to avoid information redundancy. Figure 6 B is the subset of 15 indicators that are optimized from the original 24 indicators by sorting them in descending order of feature importance scores, thus selecting the 15 indicators that contribute most to distinguishing tumors from normal samples: CD3 + CD8 + Psi - CD3 + CD8 + CD3 + CD8 + FAS - Psi - CD8 + CD45RA - CCR7 - Psi + CD8 + CD45RA - CCR7 - Psi - CD8 + FAS + Psi + CD8 + CD45RA + CCR7 + Psi + CD8 + CD45RA + CCR7 +Psi - CD8 + TOX + PD1 + CD8 + PD1 + KLRG1 - CD8 + CD45RA + CCR7 - Psi + CD8 + CD45RA + CCR7 - Psi - CD8 + CD45RA - CCR7 + Psi + CD4 + CD45RA - CCR7 + Psi - CD3 + CD4 + Psi + The top-ranked indicators are all Psi functional probes for CD8. + The diagnostic values ​​for the Tnaive subgroup are highly consistent with previous experimental data.

[0058] (5) The 15 selected parameters were reapplied to the sample (a cohort of 60 breast cancer patients diagnosed by pathological biopsy, selected according to the standards of Example 1) for 10-fold cross-validation to evaluate the stability of its performance. ROC curves were generated and AUC values ​​were calculated to comprehensively evaluate the classification ability of the model. The mean accuracy, sensitivity, specificity, and precision of 10 trials were calculated. The results are shown in the figure. When 15 biomarkers were used as diagnostic markers for breast cancer, the three-class model could accurately distinguish between Normal, Stage 1, and Stage 2 / 3 samples (Figure 6C); the two-class model could effectively distinguish between Cancer and Normal categories (Figure 6E). The accuracy, precision, sensitivity, specificity, and F1 score of both models remained at a high level of 80%–100%. Figure 6 (D, 6F) showed excellent overall performance and could effectively achieve early diagnosis of breast cancer (AUC=0.94, Figure 6G).

[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A diagnostic marker derived from peripheral blood CTL cells for breast cancer, characterized in that, Diagnostic biomarkers include one or more of the following indicators: CD3 + CD8 + Psi - CD3 + CD8 + CD3 + CD8 + FAS - Psi - CD8 + CD45RA - CCR7 - Psi + CD8 + CD45RA - CCR7 - Psi - CD8 + FAS + Psi + CD8 + CD45RA + CCR7 + Psi + CD8 + CD45RA + CCR7 + Psi - CD8 + TOX + PD1 + CD8 + KLRG1 - PD1 + CD8 + CD45RA + CCR7 - Psi + CD8 + CD45RA + CCR7 - Psi - CD8 + CD45RA - CCR7 + Psi + CD4 + CD45RA - CCR7 + Psi - or CD3 + CD4 + Psi + .

2. The application of the detection reagent for the diagnostic biomarker according to claim 1, the application includes: Application of diagnostic marker detection reagents in the preparation of breast cancer diagnostic products; or Application of diagnostic marker detection reagents in the preparation of breast cancer grading products; or Application of diagnostic marker detection reagents in the preparation of breast cancer prognostic products.

3. The application according to claim 2, characterized in that, Diagnostic biomarker detection reagents include Psi probes, the structural formula of which is: , where n is any integer from 0 to 7.

4. The application according to claim 3, characterized in that, n is 1.

5. The application according to any one of claims 2 to 4, characterized in that, The diagnostic marker detection reagents also include one or more of the following: CD3, CD4, CD8, CD45RA, CCR7, KLRG1, PD-1, FAS antibody reagent, phosphate buffer solution, or ACK lysis buffer.

6. A reagent kit for the diagnosis, grading, or prognosis of breast cancer, characterized in that, Detection reagents including the diagnostic biomarkers as described in claim 1.

7. The breast cancer diagnosis, grading, or prognosis kit according to claim 6, characterized in that, Diagnostic biomarker detection reagents include Psi probes, the structural formula of which is: , where n is any integer from 0 to 7.

8. The breast cancer diagnosis, grading, or prognosis kit according to claim 7, characterized in that, n is 1.

9. The breast cancer diagnosis, grading, or prognosis kit according to any one of claims 6 to 8, characterized in that, The diagnostic marker detection reagents also include one or more of the following: CD3, CD4, CD8, CD45RA, CCR7, KLRG1, PD-1, FAS antibody reagents, phosphate buffered saline solution, or ACK lysis buffer.

10. A system for diagnosing, grading, or prognosticating breast cancer, characterized in that, include: A biomarker acquisition module is used to acquire the diagnostic biomarker as described in claim 1; The data processing module is used to process the acquired diagnostic biomarkers, and the processing is based on the XGBoost machine learning method. The output module is used to output the diagnosis, grade, or prognosis of breast cancer.

Images