Prediction method for social distraction risk of colorectal cancer chemotherapy patient

Abstract

The invention discloses a method for predicting the social distraction risk of a colorectal cancer chemotherapy patient. The method comprises the following steps: firstly, acquiring an initial sample data set of colorectal cancer chemotherapy patients, carrying out potential profile analysis and subject working characteristic curve analysis on the patients, dividing a colorectal cancer chemotherapy patient group into a group without social distraction and a group with social distraction, and defining a social distraction threshold according to group division; on the basis, performing variable screening on an initial influence factor in the initial sample data set, determining main influence factor data and a corresponding data set, and then training a constructed initial prediction model of the social distraction risk based on the main influence factor data and the corresponding data set to obtain an initial prediction model of the social distraction risk; according to the method, a fully trained social distraction risk prediction model is obtained, and the social distraction risk of the patient is accurately predicted based on the fully trained social distraction risk prediction model, so that the risk that the colorectal cancer chemotherapy patient encounters the social distraction in the treatment process can be effectively prevented in advance.

Description

Technical Field

[0001] The present invention relates to the field of medical detection, and specifically to a method for predicting the social alienation risk of patients undergoing chemotherapy for colorectal cancer. Background Art

[0002] Colorectal cancer is a malignant tumor of the digestive tract that occurs in the colon. It is more common at the junction of the rectum and the sigmoid colon. It ranks second among gastrointestinal tumors, with the highest incidence in the 40-50 age group, and a male-female ratio of 2-3:1. Colorectal cancer is mainly adenocarcinoma, mucinous adenocarcinoma, and undifferentiated carcinoma. It can be polyp-like or ulcerative in morphology. Colorectal cancer can develop along the intestinal wall in a circular manner or infiltrate into the deep layer of the intestinal wall. In addition to metastasis through lymphatic vessels and blood flow and local invasion, it can also be implanted in the abdominal cavity or spread and metastasize along sutures and incision surfaces.

[0003] Colorectal cancer chemotherapy is an important means of treating colorectal cancer. The treatment goal is achieved by using chemical drugs to kill cancer cells. Chemotherapy drugs can enter the body through intravenous infusion or local administration, act on cancer cells, and inhibit their growth and spread. However, during chemotherapy, patients often experience anxiety and fear due to their ignorance of cancer, uncertainty about the treatment process, and possible pain and complications. The physical discomfort, lifestyle changes, and uncertainty about the future caused by chemotherapy may cause patients to feel hopeless, lost, and lonely, which in turn may lead to depression. Chemotherapy may also cause changes in the patient's appearance, such as hair loss and weight loss. These changes may affect the patient's self-esteem and self-image. Chemotherapy may even affect the patient's social and professional relationships, making it more difficult to connect with relatives, friends, and colleagues, thereby causing a serious risk of social alienation. Therefore, in order to prevent patients from experiencing this situation in advance, the existence of a method for predicting the risk of social alienation in patients undergoing colorectal cancer chemotherapy is crucial. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides a method for predicting the social alienation risk of patients undergoing chemotherapy for colon cancer, which solves the problems existing in the above-mentioned background technology.

[0005] The present invention is achieved through the following technical solutions:

[0006] A method for predicting social alienation risk in patients undergoing chemotherapy for colorectal cancer, comprising the following steps:

[0007] S1. Data collection: obtaining the initial sample data set of social alienation in patients undergoing chemotherapy for colorectal cancer. The initial sample data set includes initial influencing factor data related to the degree of social alienation, as well as social alienation data obtained by the social alienation survey tool; the social alienation data includes loneliness score data, social anxiety score data, social avoidance score data, and social alienation score data obtained by adding the three together; the initial influencing factor data related to the degree of social alienation includes the patient's demographic information and disease-related information data;

[0008] S2. Group division: by performing latent profile analysis (LPA) on the loneliness score data, social anxiety score data and social avoidance score data in the initial sample data set, the colorectal cancer chemotherapy patient group is divided into a patient group without social alienation and a patient group with social alienation;

[0009] S3. Determination of critical value: Receiver operating characteristic curve analysis was performed on the patient groups divided into those without social alienation and those with social alienation, and the threshold of social alienation was defined according to the group division;

[0010] S4, factor screening, performing variable screening on the initial influencing factor data in the initial sample data set to determine the main influencing factor data and its corresponding data set;

[0011] S5. Model construction: Using the screened main influencing factors and their corresponding data sets, a social alienation risk prediction model is established to achieve accurate prediction.

[0012] In one embodiment of the present invention, the demographic information includes: age, gender, occupation, occupational status, residential area, marital status, annual family income, medical expense payment method, education level, religious belief, Internet usage, family function, number of cohabitants and number of children.

[0013] In one embodiment of the present invention, the disease-related data include: smoking history, drinking history, body mass index, presence or absence of concurrent chronic diseases, tumor stage, diagnosis time, chemotherapy regimen, presence or absence of colostomy and symptom burden.

[0014] In one embodiment of the present invention, the group division step in step S2 is to perform latent profile analysis (LPA) on the social alienation data to divide the colorectal cancer chemotherapy patient group into a patient group with and without social alienation.

[0015] In one embodiment of the present invention, the specific steps of group division are:

[0016] The loneliness score data, social anxiety score data and social avoidance score data were used as explicit indicators, and the model was fitted starting from the initial model of category 1, and the number of potential categories was increased to 5 categories in sequence; the selection of the best model was generated by combining the interpretability of the actual clinical situation with statistical indicators, and the specific reference indicators for the selection of the best model included: Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), Sample-Size Adjusted BIC (a BIC), Bootstrap Likelihood Ratio test (BLRT), Likelihood Ratio Test (Lo-Mendell-Rubin, LMR), and entropy was used to evaluate the model classification quality, with a value range of 0 to 1, ≥0.8 indicating a good fit, and the closer the value is to 1, the more accurate the classification is.

[0017] In one embodiment of the present invention, in step S3, the threshold of social alienation is defined based on group division, including performing a receiver operating characteristic curve (ROC) analysis on the different groups of colorectal cancer chemotherapy patients, wherein the variables are social alienation score data, loneliness score data, social anxiety score data, and social avoidance score data, and calculating the corresponding area under the curve (AUC) to quantify the performance of the classifier in distinguishing different groups of colorectal cancer chemotherapy patients.

[0018] In one embodiment of the present invention, the specific operation steps for determining the critical value in step S3 are:

[0019] Data sorting and sorting: The classifier uses social alienation score data, loneliness score data, social anxiety score data, and social avoidance score data as features (i.e., input variables) to predict whether a sample belongs to a certain positive class, including the probability of each sample being predicted as a positive class and the corresponding actual label; these prediction results are sorted from high to low according to the probability value;

[0020] Threshold selection and performance evaluation: For the sorted prediction results, select each probability value as the classification threshold one by one. For each threshold, calculate the true positive rate and false positive rate. The true positive rate reflects the proportion of actual positive samples that are correctly predicted as positive, while the false positive rate reflects the proportion of actual negative samples that are incorrectly predicted as positive. By traversing all possible thresholds, a series of true positive rate and false positive rate values ​​can be obtained;

[0021] Receiver operating characteristic curve drawing: Using the true positive rate and false positive rate values ​​calculated in the above steps, draw the points corresponding to each threshold on the coordinate system with the false positive rate as the horizontal axis and the true positive rate as the vertical axis. Then, connect these points to form a smooth curve, namely the receiver operating characteristic curve. The receiver operating characteristic curve intuitively shows the performance of the classifier at different thresholds;

[0022] Calculation of area under the curve and performance quantification: Finally, the area under the receiver operating characteristic curve was calculated as a quantitative indicator of classifier performance. The larger the area under the curve, the more stable and accurate the performance of the classifier at different thresholds.

[0023] The specific calculation formula is:

[0024]

[0025] Among them, TP is the number of samples correctly predicted as positive; FN is the number of positive samples incorrectly predicted as negative samples; FP is the number of negative samples incorrectly predicted as positive samples; TN is the number of samples correctly predicted as negative samples.

[0026] In one embodiment of the present invention, in step S4, the specific steps of factor screening include:

[0027] Performing a univariate analysis on the initial sample data set to screen the initial influencing factor data and determine first influencing factor data and its corresponding first sample data set;

[0028] Applying binary logistic regression analysis to the first sample data set aims to identify and determine the main influencing factors and their corresponding main sample data sets that have a significant predictive effect on the binary results from the first influencing factor data, so as to construct an initial prediction model for the social alienation risk of colorectal cancer chemotherapy patients, and import the collected relevant data of the main influencing factors into the initial prediction model to obtain the probability of social alienation risk in colorectal cancer chemotherapy patients;

[0029] Among them, the initial prediction model for social alienation risk in patients undergoing chemotherapy for colorectal cancer is:

[0030] Logit(P)=2.949+-1.833*Religious beliefs+0.083*Symptom burden+-1.452*Good family function+-0.437*Internet usage+-2.159*Medical expenses are paid by the new rural cooperative medical insurance+-2.237*Medical expenses are paid by urban residents' medical insurance+-2.386*Medical expenses are paid by urban employees' medical insurance.

[0031] In one embodiment of the present invention, in step S5, the model building step includes:

[0032] Based on the main influencing factors and their corresponding data sets, the constructed initial prediction model of social alienation risk is trained to obtain a well-trained social alienation risk prediction model, and the patient's social alienation risk is accurately predicted based on the well-trained social alienation risk prediction model.

[0033] In one embodiment of the present invention, the evaluation index of the well-trained social alienation risk prediction model is determined based on the Bootstrap resampling method;

[0034] If the evaluation index meets the preset requirements, the patient's social alienation risk is predicted based on the fully trained social alienation risk prediction model; if the evaluation index does not meet the preset requirements, the sample size of the initial sample data set is increased, and the fully trained social alienation risk prediction model is re-determined until the evaluation index meets the preset requirements, and the patient's social alienation risk is predicted based on the fully trained social alienation risk prediction model.

[0035] The second object of the present invention is to provide a system for predicting the social alienation risk of patients undergoing chemotherapy for colon cancer, using the method for predicting the social alienation risk of patients undergoing chemotherapy for colon cancer, the system comprising:

[0036] A data collection module is used to obtain an initial sample data set of social alienation in patients undergoing chemotherapy for colorectal cancer. The initial sample data set includes initial influencing factor data related to the degree of social alienation and social alienation data obtained by a social alienation survey tool; the social alienation data includes loneliness score data, social anxiety score data, social avoidance score data, and social alienation score data obtained by adding the three together; the initial influencing factor data related to the degree of social alienation includes the patient's demographic information and disease-related information data;

[0037] A group segmentation module, which divides the colorectal cancer chemotherapy patient group into a patient group without social alienation and a patient group with social alienation by performing latent profile analysis (LPA) on the loneliness score data, social anxiety score data and social avoidance score data in the initial sample data set;

[0038] The critical value determination module conducts receiver operating characteristic curve analysis on the patient groups divided into those without social alienation and those with social alienation, and defines the threshold of social alienation based on the group division;

[0039] The factor screening module performs variable screening on the initial influencing factor data in the initial sample data set to determine the main influencing factor data and its corresponding data set;

[0040] The model building module uses the screened main influencing factors and their corresponding data sets to establish a social alienation risk prediction model to achieve accurate prediction.

[0041] Beneficial effects of the present invention:

[0042] The present invention divides the patients into a group without social alienation and a group with social alienation, thereby defining a threshold of social alienation in patients undergoing chemotherapy for colorectal cancer; and then, by constructing a social alienation risk prediction model, the probability of social alienation in the patient can be obtained, thereby effectively preventing the risk of social alienation in patients undergoing chemotherapy for colorectal cancer during the treatment process. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0044] Figure 1 It is a method flow chart of a method for predicting social alienation risk in patients undergoing chemotherapy for colorectal cancer according to the present invention;

[0045] Figure 2 It is a distribution line graph of the mean values ​​of the three dimensions of social alienation of two types of models, namely, "groups without social alienation" and "groups with social alienation" in Example 1 of the present invention;

[0046] Figure 3 is an example diagram of ROC curve data of social alienation and its three dimensions (loneliness, social anxiety and social avoidance) in Example 1 of the present invention;

[0047] Figure 4 This is the social alienation risk nomogram for colorectal cancer chemotherapy patients in Example 1 of the present invention. DETAILED DESCRIPTION

[0048] The present invention is further described below in conjunction with specific examples. These implementation cases are only used to illustrate the present invention and are not used to limit the scope of the present invention. In addition, after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms also fall within the scope limited by the claims attached to the application.

[0049] Example 1

[0050] like Figure 1-4 As shown in Tables 1 and 2, according to one aspect of the present invention, a technical solution is provided: a method for predicting social alienation risk in patients undergoing chemotherapy for colorectal cancer, comprising the following steps:

[0051] S1 Data collection: obtaining the initial sample data set of social alienation in patients undergoing chemotherapy for colorectal cancer, including social alienation data and initial influencing factor data related to the degree of social alienation;

[0052] S2 group division, clearly distinguishing between patients without social distancing and those with social distancing;

[0053] S3 critical value determination, defining the threshold of social distancing based on group division;

[0054] S4 factor screening: based on the correlation between the initial influencing factor data and the degree of social alienation of patients, the main influencing factors and corresponding data sets were selected;

[0055] The S5 model is constructed by using the screened core data set to establish a social alienation risk prediction model to achieve accurate prediction.

[0056] Compared with the prior art, the present invention first obtains the initial sample data set of patients undergoing chemotherapy for colorectal cancer, then performs potential profile analysis and receiver operating characteristic curve analysis on the patients, divides the group of patients undergoing chemotherapy for colorectal cancer into a group without social alienation and a group with social alienation, and defines the threshold of social alienation based on the group division, providing medical staff with a new classification method based on the degree of social alienation of patients. On this basis, the initial influencing factors in the initial sample data set are screened for variables, the main influencing factor data and its corresponding data set are determined, and then the constructed initial prediction model of social alienation risk is trained based on the main influencing factor data and its corresponding data set to obtain a fully trained social alienation risk prediction model, and the social alienation risk of patients is accurately predicted based on the fully trained social alienation risk prediction model, so that the risk of social alienation encountered by patients undergoing chemotherapy for colorectal cancer during treatment can be effectively prevented in advance.

[0057] In step S1, according to domestic and foreign literature on related fields, combined with the feasibility of clinical investigation and consultation with experts in related fields, the survey tool for social alienation of colorectal cancer chemotherapy patients was determined, specifically including: general information of patients (including demographic information and disease-related information, demographic information includes: age, gender, occupation, occupational status, residential area, marital status, annual family income, medical expense payment method, education level, religious belief, Internet usage, family function, number of cohabitants and number of children; disease-related information includes: smoking history, drinking history, body mass index, presence or absence of concurrent chronic diseases, tumor stage, diagnosis time, chemotherapy regimen, presence or absence of colostomy and symptom burden), social alienation survey tool (consisting of cancer loneliness scale, social anxiety scale and social avoidance scale).

[0058] It should be noted that Internet use was measured using the Internet Use Scale. The Internet Use Scale consists of 6 items, and the answer to each activity is "no" (=0) or "yes" (=1). The total score is calculated by adding these answers, ranging from 0 to 6, and the higher the score, the more frequent the Internet use. Family function is measured using the Family APGER Index Scale. The Family APGER Index Scale contains five questions, each with three possible answers: "always" gets 2 points, "sometimes" gets 1 point, and "almost never" gets 0 points. The total score of the five questions reflects different levels of family function: 7-10 points for good function, 4-6 points for moderate dysfunction, and 0-3 points for severe dysfunction. Symptom burden is measured using the Anderson Gastrointestinal Symptom Questionnaire. The Anderson Gastrointestinal Symptom Questionnaire includes 18 symptom items, and the total score is used to represent the symptom burden, thereby assessing the symptom burden. The higher the score, the heavier the symptom burden.

[0059] It should be noted that the social alienation survey tool consists of the Cancer Loneliness Scale, the Social Anxiety Scale, and the Social Avoidance Scale. The Cancer Loneliness Scale consists of 7 items, each of which is evaluated using a 5-point Likert rating method, ranging from 1 (never) to 5 (always), with a total score of 7 to 35 points. The higher the score, the higher the degree of loneliness of the cancer patient. The Social Anxiety Scale has a total of 6 items, using the Likert 5-point scoring method, 0 to 4 points represent "very inconsistent", "somewhat inconsistent", "unclear", "somewhat consistent", and "very consistent" respectively. Among them, the fourth question is a reverse scoring item, with a score range of 0 to 24 points. The higher the score, the higher the degree of social anxiety. The Social Avoidance Scale has a total of 14 items, which are scored in a "yes or no" way. Answering "yes" is 0 points and "no" is 1 point. There are 7 reverse scoring items, and the scale score range is 0 to 14 points. The higher the score, the more the patient tends to avoid and stay away from social occasions.

[0060] In step S2, the social alienation data is subjected to latent profile analysis (LPA), and based on the social alienation data (including data of three dimensions: loneliness score data, social anxiety score data and social avoidance score data), the colorectal cancer chemotherapy patient group is divided into a group without social alienation and a group with social alienation. The specific operation steps are: using the loneliness score data, social anxiety score data and social avoidance score data scores as explicit indicators, starting with category 1 (initial model) to fit the model, and increasing the number of potential categories to 5 categories in turn. The selection of the best model is generated by combining the interpretability of the actual clinical situation with statistical indicators. The reference indicators for the selection of the best model mainly include two aspects: model fitness evaluation and model comparison. Specific indicators include: Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), Sample-Size Adjusted BIC (a BIC), Bootstrap Likelihood Ratio Test (BLRT), and Lo-Mendell-Rubin (LMR). In addition, entropy is used to evaluate the classification quality of the model. The value range is 0 to 1. ≥0.8 indicates a good fit. The closer the value is to 1, the more accurate the classification. The results are shown in Tables 1 and Figure 2 ,The reference indicators selected by the best model showed that it was most reasonable to divide ,colorectal cancer chemotherapy patients into two groups: those without social alienation and those with social alienation.

[0061] Table 1 Fit index table of five models of latent profile analysis

[0062] profile AIC BIC aBIC entropy LMR(p) BLRT(p) Group Probability 1- Section 5106.408 5127.679 5108.658 2- Section 4777.992 4813.443 4781.741 0.886 <0.001 <0.001 0.54 / 0.46 3- Section 4711.095 4760.727 4716.343 0.848 0.0091 <0.001 0.45 / 0.28 / 0.27 4- Section 4656.375 4720.188 4663.123 0.871 0.0169 <0.001 0.22 / 0.12 / 0.36 / 0.30 5- Section 4635.185 4713.179 4643.433 0.867 0.6707 <0.001 0.37 / 0.14 / 0.09 / 0.12 / 0.28

[0063] In step S3, a receiver operating characteristic curve (ROC) analysis is performed on the high and low groups of patients undergoing chemotherapy for colorectal cancer, where the variables are social alienation score data, loneliness score data, social anxiety score data, and social avoidance score data, and the corresponding area under the curve (AUC) is calculated to quantify the performance of the classifier in distinguishing different groups of patients undergoing chemotherapy for colorectal cancer. The specific operation steps are: ① Data sorting and sorting: Collect the prediction results of the classifier (loneliness score data, social anxiety score data, and social avoidance score data) for the sample set, including the probability of each sample being predicted as a positive class (group with social alienation) and the corresponding actual label (group with social alienation or group without social alienation). Subsequently, these prediction results are sorted from high to low according to the probability value to provide a basis for the threshold selection in the subsequent steps. ② Threshold selection and performance evaluation: For the sorted prediction results, each probability value is selected one by one as the classification threshold. For each threshold, the true positive rate (TPR, i.e., sensitivity) and the false positive rate (FPR, i.e., 1-specificity) are calculated. TPR reflects the proportion of actual positive samples that are correctly predicted as positive, while FPR reflects the proportion of actual negative samples that are incorrectly predicted as positive. By traversing all possible thresholds, a series of TPR and FPR values ​​can be obtained. ③ROC curve drawing: Using the TPR and FPR values ​​calculated in step S3.2, with FPR as the horizontal axis and TPR as the vertical axis, draw the points corresponding to each threshold on the coordinate system. Subsequently, these points are connected to form a smooth curve, namely the receiver operating characteristic (ROC) curve. The ROC curve intuitively shows the performance of the classifier at different thresholds. ④AUC value calculation and performance quantification: Finally, the area under the ROC curve (AUC) is calculated as a quantitative indicator of classifier performance. The larger the AUC value, the more stable and accurate the performance of the classifier at different thresholds.

[0064] The specific calculation formula is:

[0065]

[0066] Among them, TP (True Positives) is the number of samples correctly predicted as positive, FN (False Negatives) is the number of positive samples incorrectly predicted as negative, FP (False Positives) is the number of negative samples incorrectly predicted as positive, and TN (True Negatives) is the number of samples correctly predicted as negative.

[0067] The results are shown in Table 2 and Figure 3, since the AUC of the social alienation score data is closest to 1, it indicates that the model of the social alienation score data is optimal. The AUC of the social alienation score data is 0.995 (95% CI: 0.990-1.000), and the optimal critical value is 37.5 based on the maximum Youden index of 0.937. Since the scale score is expressed in integers, a total score of 38 points or more indicates that patients undergoing chemotherapy for colorectal cancer are in a state of social alienation.

[0068] Table 2 Example table of ROC curve data for social isolation and its three dimensions

[0069] variable AUCs 955CI of AUCs Sensitivity Specificity Theoretical critical value Actual critical value Social distancing 0.995 0.990-1.000 0.966 0.971 37.5 38 Loneliness 0.890 0.851-0.930 0.723 0.920 20.5 21 Social anxiety 0.992 0.986-0.998 0.958 0.920 11.5 12 Social avoidance 0.864 0.819-0.910 0.874 0.745 6.5 7

[0070] Note: ROC: receiver operating characteristic; AUC: area under the ROC curve; CI: confidence interval.

[0071] In step S4, first, the collected data were statistically analyzed, and the subjects were divided into a group without social alienation and a group with social alienation according to the occurrence of outcome events. All risk factors were further analyzed univariately. If P < 0.05, the difference was statistically significant.

[0072] Secondly, the variables with significant differences were included in the multivariate logistic regression analysis. Considering the problem of logistic multicollinearity, tolerance and variance inflation factor (VIF) were used to test the multicollinearity between variables before the analysis. According to the tolerance < 0.10 and variance inflation factor > 10.0, the standard of "multicollinearity" was met. The variable screening used backward stepwise regression to establish a mathematical model of each variable and the risk of social alienation of patients. Therefore, a preliminary prediction model of the risk of social alienation in patients with colorectal cancer chemotherapy was constructed: Logit (P) = 2.949 + -1.833 * religious belief (yes) + 0.083 * symptom burden + -1.452 * family function (good) + -0.437 * Internet usage + -2.159 * medical expense payment method (new rural cooperative medical insurance) + -2.237 * medical expense payment method (urban resident medical insurance) + -2.386 * medical expense payment method (urban employee medical insurance).

[0073] In the present invention, there are five independent variables that are finally incorporated into the fully trained high social alienation risk prediction model, specifically: religious belief, symptom burden, family function, Internet usage, and medical expense payment method.

[0074] In step S5, a web-based risk calculator developed by the DynNom package in R Studio 4.2.1 software was used to draw a nomogram for the risk of social alienation in patients undergoing chemotherapy for colorectal cancer. Figure 4 shown.

[0075] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0076] The embodiments provided above are not intended to limit the scope of the present invention, and the steps described are not intended to limit the execution order thereof. Those skilled in the art may make obvious improvements to the present invention in combination with existing common knowledge, which also fall within the scope of protection defined by the claims of the present invention.

Claims

1. A method for predicting social alienation risk in patients undergoing chemotherapy for colon cancer, characterized in that: The following steps are involved: S1. Data collection: obtaining the initial sample data set of social alienation in patients undergoing chemotherapy for colorectal cancer. The initial sample data set includes initial influencing factor data related to the degree of social alienation, as well as social alienation data obtained by the social alienation survey tool; the social alienation data includes loneliness score data, social anxiety score data, social avoidance score data, and social alienation score data obtained by adding the three together; the initial influencing factor data related to the degree of social alienation includes the patient's demographic information and disease-related information data; S2. Group division: by performing latent profile analysis on loneliness score data, social anxiety score data, and social avoidance score data in the initial sample data set, the colorectal cancer chemotherapy patient group is divided into a patient group without social alienation and a patient group with social alienation; S3. Determination of critical value: Receiver operating characteristic curve analysis was performed on the patient groups divided into those without social alienation and those with social alienation, and the threshold of social alienation was defined according to the group division; S4, factor screening, performing variable screening on the initial influencing factor data in the initial sample data set to determine the main influencing factor data and its corresponding data set; S5. Model construction: Using the screened main influencing factors and their corresponding data sets, a social alienation risk prediction model is established to achieve accurate prediction.

2. The prediction method according to claim 1, characterized in that: Demographic data included: age, gender, occupation, occupational status, residential area, marital status, annual household income, medical expense payment method, educational level, religious belief, Internet usage, family function, number of cohabitants, and number of children; disease-related data included: smoking history, drinking history, body mass index, presence of concurrent chronic diseases, tumor stage, diagnosis time, chemotherapy regimen, presence of colostomy, and symptom burden.

3. The prediction method according to claim 1, characterized in that: The group division step in step S2 is to perform latent profile analysis on the social alienation data, and divide the colorectal cancer chemotherapy patient group into a patient group without social alienation and a patient group with social alienation.

4. The prediction method according to claim 3, characterized in that: The specific steps for group segmentation are as follows: The loneliness score data, social anxiety score data and social avoidance score data were used as explicit indicators, and the model was fitted starting from the initial model of category 1, and the number of potential categories was increased to 5 categories in sequence; the selection of the best model was generated by combining the interpretability of the actual clinical situation with statistical indicators. The specific reference indicators for the selection of the best model included: Aykak information criterion, Bayesian information criterion, sample-corrected BIC, Bootstrap likelihood ratio test, likelihood ratio test, and entropy was used to evaluate the model classification quality, with a value range of 0 to 1, ≥0.8 indicating a good fit, and the closer the value is to 1, the more accurate the classification.

5. The prediction method according to claim 1, characterized in that: In step S3, the threshold of social alienation is defined based on group division, including performing receiver operating characteristic curve analysis on the different groups of colorectal cancer chemotherapy patients, wherein the variables are social alienation score data, loneliness score data, social anxiety score data, and social avoidance score data, and calculating the corresponding area under the curve value to quantify the performance of the classifier in distinguishing different groups of colorectal cancer chemotherapy patients.

6. The prediction method according to claim 1, characterized in that: The specific steps for determining the critical value of step S3 are: Data sorting and sorting: The classifier uses social alienation score data, loneliness score data, social anxiety score data, and social avoidance score data as features to predict whether a sample belongs to a certain positive class, including the probability of each sample being predicted as a positive class and the corresponding actual label; these prediction results are sorted from high to low according to the probability value; Threshold selection and performance evaluation: For the sorted prediction results, select each probability value as the classification threshold one by one. For each threshold, calculate the true positive rate and false positive rate. The true positive rate reflects the proportion of actual positive samples that are correctly predicted as positive, while the false positive rate reflects the proportion of actual negative samples that are incorrectly predicted as positive. By traversing all possible thresholds, a series of true positive rate and false positive rate values ​​can be obtained; Receiver operating characteristic curve drawing: Using the true positive rate and false positive rate values ​​calculated in the above steps, draw the points corresponding to each threshold on the coordinate system with the false positive rate as the horizontal axis and the true positive rate as the vertical axis. Then, connect these points to form a smooth curve, namely the receiver operating characteristic curve. The receiver operating characteristic curve intuitively shows the performance of the classifier at different thresholds; Calculation of area under the curve and performance quantification: Finally, the area under the receiver operating characteristic curve was calculated as a quantitative indicator of classifier performance. The larger the area under the curve, the more stable and accurate the performance of the classifier at different thresholds. The specific calculation formula is: Among them, TP is the number of samples correctly predicted as positive; FN is the number of positive samples incorrectly predicted as negative samples; FP is the number of negative samples incorrectly predicted as positive samples; TN is the number of samples correctly predicted as negative samples.

7. The prediction method according to claim 1, characterized in that: In step S4, the specific steps of factor screening include: Performing a univariate analysis on the initial sample data set to screen the initial influencing factor data and determine first influencing factor data and its corresponding first sample data set; Applying binary logistic regression analysis to the first sample data set aims to identify and determine the main influencing factors and their corresponding main sample data sets that have a significant predictive effect on the binary results from the first influencing factor data, so as to construct an initial prediction model for the social alienation risk of colorectal cancer chemotherapy patients, and import the collected relevant data of the main influencing factors into the initial prediction model to obtain the probability of social alienation risk in colorectal cancer chemotherapy patients; Among them, the initial prediction model for social alienation risk in patients undergoing chemotherapy for colorectal cancer is: Logit(P)=2.949+-1.833*Religious beliefs+0.083*Symptom burden+-1.452*Good family function+-0.437*Internet usage+-2.159*Medical expenses are paid by the new rural cooperative medical insurance+-2.237*Medical expenses are paid by urban residents' medical insurance+-2.386*Medical expenses are paid by urban employees' medical insurance.

8. The prediction method according to claim 1, characterized in that: In step S5, the model building steps include: Based on the main influencing factors and their corresponding data sets, the constructed initial prediction model of social alienation risk is trained to obtain a well-trained social alienation risk prediction model, and the patient's social alienation risk is accurately predicted based on the well-trained social alienation risk prediction model.

9. The prediction method according to claim 1, characterized in that: Determining the evaluation index of the well-trained social alienation risk prediction model based on the Bootstrap resampling method; If the evaluation index meets the preset requirements, the social alienation risk of the patient is predicted based on the well-trained social alienation risk prediction model; If the evaluation index does not meet the preset requirements, the sample size of the initial sample data set is increased, and the fully trained social alienation risk prediction model is re-determined until the evaluation index meets the preset requirements, and the patient's social alienation risk is predicted based on the fully trained social alienation risk prediction model.

10. A prediction system for social alienation risk of patients undergoing chemotherapy for colon cancer, characterized in that: The method for predicting social alienation risk of a colon cancer chemotherapy patient according to any one of claims 1 to 9 is applied, wherein the system comprises: A data collection module is used to obtain an initial sample data set of social alienation in patients undergoing chemotherapy for colorectal cancer. The initial sample data set includes initial influencing factor data related to the degree of social alienation and social alienation data obtained by a social alienation survey tool; the social alienation data includes loneliness score data, social anxiety score data, social avoidance score data, and social alienation score data obtained by adding the three together; the initial influencing factor data related to the degree of social alienation includes the patient's demographic information and disease-related information data; A group segmentation module, which divides the colorectal cancer chemotherapy patient group into a patient group without social alienation and a patient group with social alienation by performing a latent profile analysis on the loneliness score data, the social anxiety score data, and the social avoidance score data in the initial sample data set; The critical value determination module conducts receiver operating characteristic curve analysis on the patient groups divided into those without social alienation and those with social alienation, and defines the threshold of social alienation based on the group division; The factor screening module performs variable screening on the initial influencing factor data in the initial sample data set to determine the main influencing factor data and its corresponding data set; The model building module uses the screened main influencing factors and their corresponding data sets to establish a social alienation risk prediction model to achieve accurate prediction.

Images