Breast cancer model construction method based on Cox regression

Abstract

The invention provides a breast cancer model construction method based on Cox regression, and relates to the technical field of clinical medical diagnosis and treatment, and the method comprises the following steps: obtaining a prognosis lifetime and a plurality of clinical factors of a breast cancer patient, associating the prognosis lifetime with the plurality of clinical factors, and constructing a sample set; performing Cox regression analysis on the basis of the sample set to obtain final clinical factors of the breast cancer patient to be predicted, and constructing a prediction model on the basis of Cox regression analysis; and performing internal verification on the prediction model based on the sample set, and obtaining the verified prediction model when the verification is passed. According to the method, the identification accuracy of the prediction model can be improved, and the prognosis survival probability of the breast cancer patient is efficiently predicted, so that reference is provided for treatment of the breast cancer patient.

Description

Technical Field

[0001] The present invention relates to the technical field of clinical medical diagnosis and treatment, and in particular to a method for constructing a breast cancer model based on Cox regression. Background Art

[0002] Breast cancer is a malignant tumor that has become a major health challenge for women worldwide. According to the global cancer data recently released by the International Agency for Research on Cancer, the number of new cases of breast cancer reached 2.3 million, accounting for 11.7% of the total new cancer cases. The incidence rate is worrying. Therefore, predicting the prognosis and survival of breast cancer patients is of great significance to improving the quality of life of breast cancer patients.

[0003] In recent years, many researchers have actively explored and applied a variety of machine learning and statistical techniques in the field of clinical research on breast cancer, aiming to build an accurate patient risk assessment prediction model. Specifically, they use strategies such as filtering, which quantifies the importance of each clinical indicator by scoring it, and then screens and eliminates factors with low scores and small impacts based on preset thresholds or quantitative standards of clinical factors; another strategy is variance selection, which measures the degree of data variation of each clinical factor and compares it with the set threshold to exclude factors with small variation and low potential contribution to the model.

[0004] However, these methods generally face a challenge when building prediction models: they often do not distinguish between differences in disease characteristics and consider all available clinical factors, resulting in a large amount of unnecessary redundant information in the model. This non-specific processing method makes it difficult to accurately capture the most critical clinical features for different disease stages or types, thereby limiting the accuracy and effectiveness of the prediction model. In view of this, there is an urgent need to develop an optimization solution that can intelligently screen and focus on the clinical factors that have the greatest impact on the prediction results based on the specific characteristics of the disease, so as to significantly improve the performance of the prediction model. Summary of the invention

[0005] The purpose of the present invention is to provide a method for constructing a breast cancer model based on Cox regression, so as to improve the problem of low accuracy of the existing prediction model due to excessive selection of clinical factors.

[0006] The present invention provides a method for constructing a breast cancer model based on Cox regression, comprising the following steps:

[0007] Obtaining the prognostic survival period and final clinical factors of breast cancer patients, and constructing a sample set after correlating the prognostic survival period with the final clinical factors;

[0008] Performing Cox regression analysis based on the sample set to obtain final clinical factors of the breast cancer patient to be predicted and constructing a prediction model based on Cox regression analysis;

[0009] The prediction model is internally verified based on the sample set, and when the verification passes, a verified prediction model is obtained.

[0010] The beneficial effect of the breast cancer model construction method based on Cox regression provided by the present invention is that the present invention uses proportional risk test, univariate Cox regression and multivariate Cox regression analysis to gradually screen out the final clinical factors with higher correlation with the prognosis and survival of breast cancer patients, reducing the influence of redundant factors on the accuracy of prediction results, and finally evaluating the performance of the prediction model from three dimensions: consistency index, calibration curve and decision analysis curve, thereby ensuring the feasibility of the model in clinical application and improving the accuracy of the prediction model.

[0011] Optionally, the breast cancer patients have a median age of 53 years and an average BMI of 23.2 kg / m 2 , 12% of the patients had a history of smoking, 55 patients had clinical pathological grade III, 22.8% of the patients received targeted drug treatment, and 25% of the patients received traditional Chinese medicine treatment. The influencing factors of clinical prognosis of breast cancer were ranked from high to low in terms of risk level: whether chemotherapy cycles were sufficient, tumor stage, histological grade, whether targeted drugs were used, and whether HE4 was positive.

[0012] Optionally, the final clinical factors are obtained by screening relevant clinical factors, and the relevant clinical factors include: the final clinical factors are obtained by screening relevant clinical factors, and the relevant clinical factors include: age, body mass index, smoking, long-term oral contraceptive experience, pregnancy, childbirth, miscarriage, menopause, combined with other diseases, family history, complications, tumor location, CA15-3, hormone receptors, human epidermal growth factor Her2 status, Ki-67 proliferation index, tumor size, lymph node metastasis, distant tumor metastasis, tumor grade, endocrine therapy, radiotherapy and radiotherapy cycles, chemotherapy and chemotherapy cycles, targeted therapy, traditional Chinese medicine treatment, the final clinical factors are: age, menopause, CA15-3, Ki-67 proliferation index, tumor size, lymph node metastasis, distant tumor metastasis, tumor grade, endocrine therapy, radiotherapy and radiotherapy cycles, chemotherapy and chemotherapy cycles, targeted therapy.

[0013] Optionally, the process of constructing the prediction model includes a Cox regression analysis process; in the Cox regression process, the optimal parameter values ​​are determined by Cox regression fitting, and the prediction model is constructed based on the coefficient values.

[0014] Optionally, when internal verification is performed based on the sample set, it includes:

[0015] Based on the prediction model, the final clinical factors in the sample set are internally verified to obtain the prediction results of the breast cancer patients to be internally verified, and after comparing the prediction results of the breast cancer patients to be internally verified with the corresponding actual results in the test sample set, the similarity score, calibration curve, and decision analysis curve between the prediction results of the breast cancer patients to be internally verified and the actual results are obtained; when the similarity score is greater than the excellent score threshold, the calibration curves of the predicted disease probability and the actual disease probability in the internal verification sample set fit each other, and the net benefit of the internal verification sample set is greater than the excellent net benefit threshold, the internal verification is passed, otherwise the number of initial sample sets is increased and iteration continues.

[0016] Optionally, among the relevant clinical factors, age, BMI, tumor size, and lymph node metastasis are quantified as 1 / 2 / 3 / 4, pregnancy, Ki-67 proliferation index, and tumor grade are re-quantified as 1 / 2 / 3, and the remaining relevant clinical factors are quantified as 0 / 1.

[0017] Optionally, the breast cancer patients are divided into patients who have undergone breast-conserving surgery and patients who have not undergone breast-conserving surgery. If the patients undergo breast-conserving surgery, they need to receive postoperative radiotherapy. Among them, breast cancer patients who need adjuvant radiotherapy must meet the following conditions:

[0018] The maximum diameter of the primary tumor is >5 cm, or the tumor invades the breast skin or chest wall;

[0019] ≥4 axillary lymph node metastases, or supraclavicular or internal mammary lymph node metastases;

[0020] Patients with primary tumor stage T1-2 and 1-3 axillary lymph node metastases.

[0021] Optionally, adjuvant radiotherapy can be omitted for breast cancer patients who are ≥50 years old, have tumor grade I-II, no vascular tumor thrombus, have one axillary lymph node metastasis, and are positive for estrogen receptor ER and progesterone receptor PR.

[0022] The inclusion criteria for breast cancer patients were as follows: patients diagnosed with primary breast cancer by histopathology and cytopathology; patients with complete medical records; patients with complete and available staging and grading information; patients receiving regular postoperative adjuvant radiotherapy or chemotherapy;

[0023] The exclusion criteria for breast cancer patients were as follows: missing baseline data of patients, unclear clinical stage and grade; lack of some clinical or laboratory diagnostic information such as lymph node status, surgical and chemoradiotherapy treatment measures, estrogen receptor ER, progesterone receptor PR, human epidermal growth factor HER2 information; patients with tumors in other organs and tissues; patients with a history of other surgeries; patients with complications such as severe heart, liver, and kidney lesions and incomplete clinical data; patients who did not receive surgical treatment; no follow-up records; patients died of other causes.

[0024] Optionally, the final clinical factors are divided into the following categories according to the degree of risk from high to low: whether chemotherapy cycles are sufficient, tumor stage, histological grade, whether targeted drugs are used, and whether HE4 is positive. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a basic step flow chart of a method for building a breast cancer model based on Cox regression provided by the present invention;

[0026] Figure 2 It is a nomogram of a breast cancer model construction method based on Cox regression provided by the present invention;

[0027] Figure 3 It is a consistency index curve diagram of the internal verification of the sample set provided by the present invention;

[0028] Figure 4 is a calibration curve diagram for internal verification of a sample set provided by the present invention;

[0029] Figure 5 It is a decision analysis curve diagram of internal verification of the sample set provided by the present invention. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention. Unless otherwise defined, the technical terms or scientific terms used herein should be understood by people with general skills in the field to which the present invention belongs. "Including" and similar words used in this article mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects.

[0031] The present invention provides a method for constructing a breast cancer model based on Cox regression. Figure 1 , the method comprises the following steps:

[0032] S1. Obtain the prognosis survival period and final clinical factors of breast cancer patients, and construct a sample set after correlating the prognosis survival period with the final clinical factors;

[0033] S2. Performing Cox regression analysis based on the sample set to obtain the final clinical factors of the breast cancer patient to be predicted and constructing a prediction model based on Cox regression analysis;

[0034] S3. The prediction model is internally verified based on the sample set, and when the verification passes, a verified prediction model is obtained.

[0035] In some embodiments, when executing step S1, the process of constructing the sample set includes:

[0036] According to the inclusion and exclusion criteria, a total of 246 patients diagnosed with breast cancer in Jingdezhen Second People's Hospital from January 2016 to December 2023 were selected as the sample set for inclusion in the study.

[0037] Specifically, the median age of breast cancer patients was 53 years old, and the average BMI was 23.2 kg / m 2 , 12% of the patients had a history of smoking, 55 patients had clinical pathological grade III, 22.8% of the patients received targeted drug treatment, and 25% of the patients received traditional Chinese medicine treatment. The influencing factors of clinical prognosis of breast cancer were ranked from high to low in terms of risk level: whether chemotherapy cycles were sufficient, tumor stage, histological grade, whether targeted drugs were used, and whether HE4 was positive.

[0038] Furthermore, when breast cancer patients were included in the study, the inclusion and exclusion criteria had to be met.

[0039] Specifically, the inclusion criteria must meet the following: (1) patients diagnosed with primary breast cancer by histopathology and cytopathology. (2) patients with complete medical records. (3) patients with complete and available staging and grading information. (4) patients receiving postoperative adjuvant regular radiotherapy or chemotherapy.

[0040] Furthermore, the exclusion criteria must meet the following: (1) Patients with missing baseline data and unclear clinical stage and grade. (2) Patients lacking some clinical or laboratory diagnostic information such as lymph node status, surgical and chemoradiotherapy treatment measures, estrogen receptor ER, progesterone receptor PR, and human epidermal growth factor HER2 information. (3) Patients with tumors of other organs and tissues. (4) Patients with a history of other surgeries. (5) Patients with complications such as other severe temperamental lesions such as heart, liver, and kidney, and patients with incomplete clinical data. (6) Patients who did not undergo surgical treatment. (7) No follow-up records. (8) Patients who died of other causes or whose cause of death is unknown.

[0041] In some embodiments, when executing step S1, the process of obtaining the final clinical factors of the breast cancer patient includes:

[0042] S1-1. Obtain relevant clinical factors of breast cancer patients;

[0043] S1-2, setting a threshold probability of the relevant clinical factors and the prognosis survival period, and screening the relevant clinical factors based on the threshold probability and univariate Cox regression to obtain preliminary screening clinical factors;

[0044] S1-3, performing probability parameter standard analysis and screening on the preliminary screening clinical factors based on the proportional risk test to obtain preset clinical factors;

[0045] S1-4. Perform independence test on the preset clinical factors based on multivariate Cox regression to obtain the final clinical factors.

[0046] Specifically, the relevant clinical factors in step S1-1 refer to the factors that may affect the prognosis and survival of breast cancer patients clinically. The final clinical factors are obtained by screening the relevant clinical factors, and the relevant clinical factors include: age, body mass index, smoking, long-term oral contraceptive experience, pregnancy, childbirth, miscarriage, menopause, combined with other diseases, family history, complications, tumor location, CA15-3, hormone receptors, human epidermal growth factor Her2 status, Ki-67 proliferation index, tumor size, lymph node metastasis, distant tumor metastasis, tumor grade, endocrine therapy, radiotherapy and radiotherapy cycle, chemotherapy and chemotherapy cycle, targeted therapy, and traditional Chinese medicine treatment.

[0047] In fact, when executing step S1-1, breast cancer patients can be divided into patients who need adjuvant radiotherapy and those who do not need adjuvant radiotherapy. If the breast cancer patients who need adjuvant radiotherapy are ≥50 years old, have tumor grade I-II, have no vascular tumor thrombus, have one axillary lymph node metastasis, and are estrogen receptor ER and progesterone receptor PR positive, adjuvant radiotherapy can be omitted.

[0048] Furthermore, the final clinical factors are: age, menopause, CA15-3, Ki-67 proliferation index, tumor size, lymph node metastasis, distant tumor metastasis, tumor grade, endocrine therapy, radiotherapy and radiotherapy cycle, chemotherapy and chemotherapy cycle, targeted therapy. Table 1 shows the quantification scheme of relevant clinical factors.

[0049] Table 1 Quantification scheme of relevant clinical factors

[0050]

[0051]

[0052] Specifically, among the relevant clinical factors, age, BMI, tumor size, and lymph node metastasis were quantified as 1 / 2 / 3 / 4, pregnancy, Ki-67 proliferation index, and tumor grade were re-quantified as 1 / 2 / 3, and the remaining relevant clinical factors were quantified as 0 / 1.

[0053] Furthermore, breast cancer patients are divided into those who receive breast-conserving surgery and those who do not. If they receive breast-conserving surgery, they need to receive postoperative radiotherapy. Among them, breast cancer patients who need adjuvant radiotherapy must meet the following conditions:

[0054] 1) The maximum diameter of the primary tumor is >5 cm, or the tumor invades the breast skin or chest wall;

[0055] 2) ≥4 axillary lymph node metastases, or supraclavicular or internal mammary lymph node metastases;

[0056] 3) Patients with primary tumor stage T1-2 and 1-3 axillary lymph node metastases.

[0057] Specifically, the final clinical factors are divided into the following categories according to the degree of risk: whether the chemotherapy cycle is sufficient, tumor stage, histological grade, whether targeted drugs are used, and whether HE4 is positive.

[0058] Specifically, when executing step S1-2, the steps for obtaining the initial screening clinical factors are as follows:

[0059] The threshold probability of the initial inclusion of clinical factors was set as α=10%, and the P value of each relevant clinical factor was calculated. The relevant clinical factors corresponding to P≤α were screened out from the relevant clinical factors as the initial screening clinical factors; the initial screening clinical factors were: age, menopause, CA15-3, hormone receptors, human epidermal growth factor Her2 status, Ki-67 proliferation index, tumor size, lymph node metastasis, distant tumor metastasis, tumor grade, endocrine therapy, radiotherapy and radiotherapy cycles, chemotherapy and chemotherapy cycles, targeted therapy. Table 2 shows the quantitative scheme of the initial screening clinical factors.

[0060] Table 2 Quantification scheme of clinical factors for initial screening

[0061]

[0062] Furthermore, in step S1-3, the process of performing probability parameter standard analysis and screening on the primary screening clinical factors to obtain the preset clinical factors includes:

[0063] After performing a proportional risk test on the initial clinical factors, the P value of each initial screening clinical factor was calculated. The initial screening clinical factors hormone receptor (EP / PR) and Her2 status with P < 0.05 were eliminated, and the remaining 12 factors were the preset clinical factors.

[0064] Specifically, the preset clinical factors are: age, menopause, CA15-3, Ki-67 value-added index, tumor size, lymph node metastasis, distant tumor metastasis, tumor grade, endocrine therapy, radiotherapy and radiotherapy cycle, chemotherapy and chemotherapy cycle, and targeted therapy.

[0065] Furthermore, the process of obtaining the final clinical factors in step S1-4 includes:

[0066] Based on multivariate Cox regression, the independence test of the preset clinical factors was performed, and the final clinical factors were obtained after eliminating the associated preset clinical factors. Since each preset clinical factor is independent of each other, the final clinical factors are still: age, menopause, CA15-3, Ki-67 value-added index, tumor size, lymph node metastasis, distant tumor metastasis, tumor grade, endocrine therapy, radiotherapy and radiotherapy cycles, chemotherapy and chemotherapy cycles, targeted therapy. Table 3 shows the quantification scheme of the final clinical factors.

[0067] Table 3 Quantification scheme of final clinical factors

[0068]

[0069] In fact, when the independence test of the preset clinical factors is performed based on multivariate Cox regression and the final clinical factors are obtained after eliminating the correlated preset clinical factors, when two or more preset clinical factors have a functional relationship or a linear relationship, any one of the preset clinical factors should be selected as an independent variable and added to the set of final clinical factors.

[0070] In some embodiments, when executing step S3, the prediction model performs internal verification based on the sample set, including:

[0071] Based on the prediction model fitted by Cox regression, the final clinical factors in the sample set are internally verified to obtain the internal verification results of the breast cancer patients to be internally verified, and after comparing the internal verification results of the breast cancer patients to be internally verified with the corresponding true results in the sample set, a similarity score curve, a calibration curve, and a decision analysis curve are obtained between the internal verification results of the breast cancer patients to be internally verified and the true results;

[0072] When the similarity score is greater than the excellent score threshold, the calibration curves of the predicted disease probability and the actual disease probability in the sample set fit each other, and the net benefit of the sample set is greater than the excellent net benefit threshold, the internal validation passes, otherwise the number of sample sets is increased and iterations continue.

[0073] Specifically, the consistency index is used to estimate the similarity between the actual result score and the predicted result score ranking. Its theoretical value is between 0 and 1. When the consistency index is > 0.5, it means that the prediction model with good Cox regression fitting has better performance. The calculated consistency index curve of the internal validation sample set is as follows: Figure 3 As shown, Figure 3 The consistency index curves corresponding to 12 variables and 11 variables are displayed. The results show that the consistency index of the sample set is greater than 0.5, which means that the prediction model fitted by Cox regression fits the sample set well.

[0074] Furthermore, the calibration curves of the sample set and the internal validation sample set are as follows Figure 4 As shown in the figure, the functional relationship between the survival proportion and the generated prediction value is expressed. The survival proportion is divided into 24-month survival proportion, 36-month survival proportion, 60-month survival proportion and 90-month survival proportion. The survival prediction includes 24-month survival prediction, 36-month survival prediction, 60-month survival prediction and 90-month survival prediction. The results show that the trained prediction model shows good calibration performance and can more accurately predict the actual survival rates of one, two, three and five years.

[0075] Furthermore, Figure 5 The decision analysis curves of the training sample set and the internal test set are shown, indicating the functional relationship between the net benefit and the threshold probability. The independent variables include: the net benefit of all treatments, the net benefit of no treatment, the smoothed net benefit of the validation set, and the smoothed net benefit of the training set. Figure 5 It can be seen that the personal threshold probability of breast cancer patients is 60% (that is, the probability of breast cancer patients surviving for 3 years is 60%, then breast cancer patients will choose advice), and the net benefit is 0.145, which indicates that the patient should choose further treatment.

[0076] Although the embodiments of the present invention are described in detail above, it is obvious to those skilled in the art that various modifications and variations can be made to these embodiments. However, it should be understood that such modifications and variations are within the scope and spirit of the present invention as described in the claims. Moreover, the present invention described herein may have other embodiments and may be implemented or realized in a variety of ways.

Claims

1. A method for constructing a breast cancer model based on Cox regression, characterized in that: include: Obtaining the prognostic survival period and final clinical factors of breast cancer patients, and constructing a sample set after correlating the prognostic survival period with the final clinical factors; Performing Cox regression analysis based on the sample set to obtain final clinical factors of the breast cancer patient to be predicted and constructing a prediction model based on Cox regression analysis; The prediction model is internally verified based on the sample set, and when the verification passes, a verified prediction model is obtained.

2. A method for constructing a breast cancer model based on Cox regression according to claim 1, characterized in that: The median age of the breast cancer patients was 53 years old and the average BMI was 23.2 kg / m 2 , 12% of the patients had a history of smoking, 55 patients had clinical pathological grade III, 22.8% of the patients received targeted drug treatment, and 25% of the patients received traditional Chinese medicine treatment. The influencing factors of clinical prognosis of breast cancer were ranked from high to low in terms of risk level: whether chemotherapy cycles were sufficient, tumor stage, histological grade, whether targeted drugs were used, and whether HE4 was positive.

3. The method for constructing a breast cancer model based on Cox regression according to claim 1, characterized in that: The final clinical factors were obtained by screening relevant clinical factors, including: age, body mass index, smoking, long-term oral contraceptive experience, pregnancy, childbirth, miscarriage, menopause, combined with other diseases, family history, complications, tumor location, CA15-3, hormone receptors, human epidermal growth factor Her2 status, Ki-67 value-added index, tumor size, lymph node metastasis, distant tumor metastasis, tumor grade, endocrine therapy, radiotherapy and radiotherapy cycles, chemotherapy and chemotherapy cycles, targeted therapy, and traditional Chinese medicine treatment. The final clinical factors were: age, menopause, CA15-3, Ki-67 value-added index, tumor size, lymph node metastasis, distant tumor metastasis, tumor grade, endocrine therapy, radiotherapy and radiotherapy cycles, chemotherapy and chemotherapy cycles, and targeted therapy.

4. The method for constructing a breast cancer model based on Cox regression according to claim 1, characterized in that: The process of constructing the prediction model includes a Cox regression analysis process; in the Cox regression process, the optimal parameter value is determined by Cox regression fitting, and the prediction model is constructed based on the coefficient value.

5. The method for constructing a breast cancer model based on Cox regression according to claim 1, characterized in that: When internal validation is performed based on the sample set, it includes: Based on the prediction model, the final clinical factors in the sample set are internally verified to obtain the prediction results of the breast cancer patients to be internally verified, and after comparing the prediction results of the breast cancer patients to be internally verified with the corresponding actual results in the test sample set, the similarity score, calibration curve, and decision analysis curve between the prediction results of the breast cancer patients to be internally verified and the actual results are obtained; when the similarity score is greater than the excellent score threshold, the calibration curves of the predicted disease probability and the actual disease probability in the internal verification sample set fit each other, and the net benefit of the internal verification sample set is greater than the excellent net benefit threshold, the internal verification is passed, otherwise the number of initial sample sets is increased and iteration continues.

6. The method for constructing a breast cancer model based on Cox regression according to claim 3, characterized in that: Among the relevant clinical factors, age, BMI, tumor size, and lymph node metastasis were quantified as 1 / 2 / 3 / 4, pregnancy, Ki-67 proliferation index, and tumor grade were re-quantified as 1 / 2 / 3, and the remaining relevant clinical factors were quantified as 0 / 1.

7. The method for constructing a breast cancer model based on Cox regression according to claim 1, characterized in that: The breast cancer patients are divided into patients who have undergone breast-conserving surgery and patients who have not undergone breast-conserving surgery. If patients undergo breast-conserving surgery, they need to receive postoperative radiotherapy. Among them, breast cancer patients who need adjuvant radiotherapy must meet the following conditions: The maximum diameter of the primary tumor is >5 cm, or the tumor invades the breast skin or chest wall; ≥4 axillary lymph node metastases, or supraclavicular or internal mammary lymph node metastases; Patients with primary tumor stage T1-2 and 1-3 axillary lymph node metastases.

8. The method for constructing a breast cancer model based on Cox regression according to claim 1, characterized in that: Adjuvant radiotherapy can be omitted if the breast cancer patient is ≥50 years old, has a tumor grade of I to II, has no vascular tumor thrombus, has one axillary lymph node metastasis, and is positive for estrogen receptor ER and progesterone receptor PR.

9. The method for constructing a breast cancer model based on Cox regression according to claim 1, characterized in that: The inclusion criteria for breast cancer patients were as follows: patients diagnosed with primary breast cancer by histopathology and cytopathology; patients with complete medical records; patients with complete and available staging and grading information; patients receiving regular postoperative adjuvant radiotherapy or chemotherapy; The exclusion criteria for breast cancer patients were as follows: missing baseline data of patients, unclear clinical stage and grade; lack of some clinical or laboratory diagnostic information such as lymph node status, surgical and chemoradiotherapy treatment measures, estrogen receptor ER, progesterone receptor PR, human epidermal growth factor HER2 information; patients with tumors in other organs and tissues; patients with a history of other surgeries; patients with complications such as severe heart, liver, and kidney lesions and incomplete clinical data; patients who did not receive surgical treatment; no follow-up records; patients died of other causes.

10. The method for constructing a breast cancer model based on Cox regression according to claim 1, characterized in that: The final clinical factors are divided into the following categories according to the degree of risk: whether the chemotherapy cycle is sufficient, tumor stage, histological grade, whether targeted drugs are used, and whether HE4 is positive.

Images