Prediction method and system for occurrence of anastomotic fistula after colorectal cancer operation

Abstract

The invention belongs to the technical field of medicine, and discloses a method and system for predicting anastomotic fistula occurring after colorectal cancer operation, and the method comprises the steps: obtaining clinical pathology data and laboratory analysis data of a patient; filling missing data in the clinical pathology data and the laboratory analysis data; screening key indexes by using a single-factor regression analysis technology; performing multi-factor regression analysis on the key indexes, and screening to obtain independent prediction factors for predicting whether the colorectal cancer patient has the anastomotic fistula or not after operation; based on the independent prediction factors obtained through screening, an anastomotic fistula prediction model is constructed and trained, and the trained anastomotic fistula prediction model is used for predicting the postoperative anastomotic fistula occurrence probability of the colorectal cancer patient. The method can more accurately predict the probability of occurrence of anastomotic fistula of a colorectal cancer patient after an operation, has better accuracy and reliability, can facilitate preoperative prediction of the probability of occurrence of anastomotic fistula for clinical doctors, and assists in guiding treatment decision.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular to a method and system for predicting anastomotic leakage after colorectal cancer surgery. Background Art

[0002] Colorectal cancer (CRC) is one of the most common cancers of the digestive system. Although surgical techniques for colorectal cancer have improved significantly over the past few decades, postoperative complications still plague surgeons. Anastomotic leakage (AL) is one of the most serious complications after CRC surgery, which has an adverse effect on the patient's survival rate and quality of life, and also leads to prolonged hospitalization and increased medical costs.

[0003] However, how to accurately predict anastomotic fistula after colorectal cancer surgery is a clinical problem. Clinically, doctors often judge the probability of occurrence of patients based on experience to formulate diagnosis and treatment plans. However, when explaining the patient's condition to the patient's family, an objective and quantifiable judgment probability is often not given. Therefore, a system that can accurately and objectively predict anastomotic fistula after colorectal cancer surgery is needed to assist clinicians, and how to establish a prediction model for this system that predicts anastomotic fistula after colorectal cancer surgery is the core technology.

[0004] Computed tomography (CT) is a commonly used radiological method in medical imaging. It can obtain thin cross-sectional high-resolution images, which can be used to quantify the content of fat and muscle tissue. Compared with body mass index (BMI), these parameters can fully reflect the detailed information of body composition. Although almost all colorectal cancer patients undergo abdominal CT examination before surgery, the evaluation of CT examination is limited to evaluating the tumor stage and the presence of distant metastasis. The value of CT images for body composition has not yet been applied in clinical applications.

[0005] Therefore, how to provide a prediction method and system for anastomotic leakage after colorectal cancer surgery is an urgent problem to be solved. Summary of the invention

[0006] The embodiment of the present invention provides a method and system for predicting anastomotic leakage after colorectal cancer surgery to solve the above-mentioned technical problems in the prior art.

[0007] In order to have a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended to be a general review, nor is it intended to identify key / important components or to delineate the scope of protection of these embodiments. Its only purpose is to present some concepts in a simple form as a preface to the detailed description that follows.

[0008] According to a first aspect of an embodiment of the present invention, a method for predicting anastomotic leakage after colorectal cancer surgery is provided.

[0009] In one embodiment, the method for predicting anastomotic leakage after rectal cancer surgery comprises:

[0010] Obtain the patient's clinical pathological data and laboratory analysis data, and outline the target area on the L3 plane to obtain the skeletal muscle area, visceral fat area, subcutaneous fat area, and intermuscular fat area on the L3 plane;

[0011] Based on the random forest chain equation multiple imputation algorithm, missing data in clinical pathology data and laboratory analysis data were filled;

[0012] The variables were screened by univariate regression analysis technology to obtain the key indicators for analyzing anastomotic leakage after colorectal cancer surgery; linear tests, multicollinearity tests and strong influence point analyses were performed on the key indicators;

[0013] Based on multivariate regression analysis technology, multivariate regression analysis was performed on key indicators with significant probability values ​​less than the preset threshold, and independent predictive factors for predicting whether anastomotic leakage would occur in patients with colorectal cancer after surgery were screened;

[0014] Based on the independent predictive factors screened, an anastomotic fistula prediction model was constructed and trained, and the trained anastomotic fistula prediction model was used to predict the probability of anastomotic fistula after surgery in patients with colorectal cancer.

[0015] In one embodiment, the clinical pathological data include age, gender, body mass index, history of hypertension, history of coronary heart disease, history of smoking, history of drinking, NRS2002 score, tumor T stage, tumor N stage, tumor differentiation degree, surgical method and surgical site;

[0016] The laboratory analysis data included preoperative neutrophil count, lymphocyte count, white blood cell count, platelet count, albumin level, alanine aminotransferase level, aspartate aminotransferase level, activated partial thromboplastin time, antithrombin III activity, fasting blood glucose level, and neutrophil-to-lymphocyte ratio.

[0017] In one embodiment, when outlining the skeletal muscle area in the L3 plane, the CT value range is set to -29 to 150 HU;

[0018] When outlining the visceral fat area in the L3 plane, the CT value range was set to -150 to -50 HU;

[0019] When outlining the subcutaneous fat area and intermuscular fat area in the L3 plane, the CT value range was set to -190 to -30 HU.

[0020] In one embodiment, the key indicators include gender, white blood cell count, albumin level, fasting blood glucose level, NRS2002 score, skeletal muscle area at lumbar 3 plane, and visceral fat area at lumbar 3 plane;

[0021] The independent predictive factors included gender, fasting blood glucose level, NRS2002 score, skeletal muscle area at L3 plane and visceral fat area at L3 plane.

[0022] In one embodiment, the performing of linearity test, multicollinearity test and strong influence point analysis on key indicators respectively includes:

[0023] Based on the natural logarithm formula, the natural logarithm values ​​of the continuous indicators in the key indicators were calculated, and the Box-Tidwell method was used to test whether there was a linear relationship between the independent variables and the logistic transformation values ​​of the dependent variables;

[0024] Use tolerance or variance inflation factors to diagnose multicollinearity among independent variables;

[0025] Based on the comparison of each patient's Cook's distance with the preset distance threshold, we analyzed whether there were strong influencing points in the data set.

[0026] According to a second aspect of an embodiment of the present invention, a system for predicting anastomotic leakage after colorectal cancer surgery is provided.

[0027] In one embodiment, the prediction system for anastomotic leakage after colorectal cancer surgery comprises:

[0028] The clinical characteristic data acquisition module is used to obtain the patient's clinical pathological data and laboratory analysis data, and to outline the target area on the L3 plane to obtain the skeletal muscle area, visceral fat area, subcutaneous fat area and intermuscular fat area on the L3 plane;

[0029] Clinical feature data processing module, used to fill in missing data in clinical pathology data and laboratory analysis data based on random forest chain equation multiple imputation algorithm;

[0030] The univariate regression analysis screening module is used to screen variables using univariate regression analysis technology to obtain key indicators for analyzing anastomotic leakage after colorectal cancer surgery; and to perform linear tests, multicollinearity tests, and strong influence point analyses on the key indicators;

[0031] A multi-factor regression analysis screening module is used to perform multi-factor regression analysis on key indicators with significance probability values ​​less than a preset threshold based on multi-factor regression analysis technology, and screen out independent predictive factors for predicting whether anastomotic fistula will occur in patients with colorectal cancer after surgery;

[0032] The anastomotic fistula occurrence probability prediction module is used to construct and train an anastomotic fistula prediction model based on the independent predictive factors obtained by screening, and use the trained anastomotic fistula prediction model to predict the probability of anastomotic fistula occurring in colorectal cancer patients after surgery.

[0033] In one embodiment, the clinical pathological data include age, gender, body mass index, history of hypertension, history of coronary heart disease, history of smoking, history of drinking, NRS2002 score, tumor T stage, tumor N stage, tumor differentiation degree, surgical method and surgical site;

[0034] The laboratory analysis data included preoperative neutrophil count, lymphocyte count, white blood cell count, platelet count, albumin level, alanine aminotransferase level, aspartate aminotransferase level, activated partial thromboplastin time, antithrombin III activity, fasting blood glucose level, and neutrophil-to-lymphocyte ratio.

[0035] In one embodiment, when the clinical characteristic data acquisition module outlines the target area in the L3 plane, the CT value range when outlining the skeletal muscle area in the L3 plane is set to -29 to 150HU; the CT value range when outlining the visceral fat area in the L3 plane is set to -150 to -50HU; and the CT value range when outlining the subcutaneous fat area and intermuscular fat area in the L3 plane is set to -190 to -30HU.

[0036] In one embodiment, the key indicators include gender, white blood cell count, albumin level, fasting blood glucose level, NRS2002 score, skeletal muscle area at lumbar 3 plane, and visceral fat area at lumbar 3 plane;

[0037] The independent predictive factors included gender, fasting blood glucose level, NRS2002 score, skeletal muscle area at L3 plane and visceral fat area at L3 plane.

[0038] In one embodiment, when the univariate regression analysis screening module performs linear test, multicollinearity test and strong influence point analysis on the key indicators respectively, the natural logarithm formula is used to calculate the natural logarithm value of the continuous indicator in the key indicator, and the Box-Tidwell method is used to test whether there is a linear relationship between the independent variable and the logistic transformation value of the dependent variable; tolerance or variance inflation factor is used to diagnose multicollinearity between independent variables; based on the comparison results of the Cook's distance of each patient with the preset distance threshold, whether there are strong influence points in the data set is analyzed.

[0039] According to a third aspect of an embodiment of the present invention, a computer device is provided.

[0040] In some embodiments, the computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0041] According to a fourth aspect of embodiments of the present invention, a computer-readable storage medium is provided.

[0042] In one embodiment, the computer-readable storage medium stores a computer program, and the computer program implements the steps of the above method when executed by a processor.

[0043] The technical solution provided by the embodiment of the present invention may have the following beneficial effects:

[0044] The present invention uses 5 clinical indicators and imaging indicators that are significantly correlated with the outcomes to construct a prediction model for predicting the occurrence of anastomotic fistula after surgery in patients with colorectal cancer, and optimizes the coefficients in the model equation to obtain a suitable model, which can more accurately predict the probability of anastomotic fistula after surgery in patients with colorectal cancer, with an AUC value of up to 0.848, which has good accuracy and reliability, and can facilitate clinicians to predict the probability of anastomotic fistula before surgery, and assist in guiding treatment decisions.

[0045] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0047] Figure 1 is a flow chart of a method for predicting anastomotic leakage after colorectal cancer surgery according to an exemplary embodiment;

[0048] Figure 2 is a structural block diagram of a prediction system for anastomotic leakage after colorectal cancer surgery according to an exemplary embodiment;

[0049] Figure 3 is a schematic diagram showing visualization of a fitted logistic regression model as a nomogram according to an exemplary embodiment;

[0050] Figure 4 is a schematic diagram of a receiver operating characteristic curve and the area under the curve of a prediction result obtained by a prediction model according to an exemplary embodiment;

[0051] Figure 5 is a schematic diagram of a calibration curve of a prediction result obtained by a prediction model according to an exemplary embodiment;

[0052] Figure 6 The figure is a schematic diagram showing the structure of a computer device according to an exemplary embodiment. DETAILED DESCRIPTION

[0053] The following description and accompanying drawings fully illustrate the specific embodiments of this article so that those skilled in the art can practice them. Parts and features of some embodiments may be included in or replace parts and features of other embodiments. The scope of the embodiments of this article includes the entire scope of the claims, as well as all available equivalents of the claims. Herein, the terms "first", "second", etc. are only used to distinguish one element from another, without requiring or implying any actual relationship or order between these elements. In fact, the first element can also be called the second element, and vice versa. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that the structure, device or equipment including a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also include elements inherent to such structure, device or equipment. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the structure, device or equipment including the elements. Each embodiment is described in a progressive manner herein, and each embodiment focuses on the differences from other embodiments, and the same and similar parts between the embodiments can be referred to each other.

[0054] The terms "longitudinal", "lateral", "upper", "lower", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc. in this document indicate the orientation or position relationship based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing this document and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as a limitation on the present invention. In the description of this document, unless otherwise specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a mechanical connection or an electrical connection, it can also be the internal communication of two elements, it can be a direct connection, or it can be an indirect connection through an intermediate medium. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0055] As used herein, the term "plurality" means two or more than two, unless otherwise specified.

[0056] In this document, the character " / " indicates that the preceding and following objects are in an "or" relationship. For example, A / B means: A or B.

[0057] In this article, the term "and / or" is a description of the association relationship between objects, indicating that three relationships may exist. For example, A and / or B means: A or B, or, A and B.

[0058] It should be understood that, although the various steps in the flow chart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the figure may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a portion of the sub-steps or stages of other steps.

[0059] Each module in the device or system of the present application can be implemented in whole or in part by software, hardware, or a combination thereof. The above modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to the above modules.

[0060] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0061] Figure 1 An embodiment of a method for predicting anastomotic leakage after colorectal cancer surgery according to the present invention is shown.

[0062] In this optional embodiment, the method for predicting anastomotic leakage after colorectal cancer surgery includes:

[0063] S101, obtaining the patient's clinical pathological data and laboratory analysis data, and outlining the target area on the L3 plane to obtain the skeletal muscle area, visceral fat area, subcutaneous fat area, and intermuscular fat area on the L3 plane;

[0064] S103, filling missing data in clinical pathology data and laboratory analysis data based on random forest chain equation multiple interpolation algorithm;

[0065] S105. Filter variables using univariate regression analysis technology to obtain key indicators for analyzing anastomotic leakage after colorectal cancer surgery; and perform linear tests, multicollinearity tests, and strong influence point analyses on the key indicators;

[0066] S107. Based on multivariate regression analysis technology, multivariate regression analysis was performed on key indicators with significant probability values ​​less than the preset threshold, and independent predictive factors for predicting whether anastomotic leakage would occur in patients with colorectal cancer after surgery were screened;

[0067] S109. Based on the independent predictive factors obtained through screening, a prediction model for anastomotic fistula is constructed and trained, and the trained prediction model for anastomotic fistula is used to predict the probability of anastomotic fistula occurring after surgery in patients with colorectal cancer.

[0068] Figure 2 An embodiment of a system for predicting anastomotic leakage after colorectal cancer surgery according to the present invention is shown.

[0069] In this optional embodiment, the prediction system for anastomotic leakage after colorectal cancer surgery includes:

[0070] The clinical characteristic data acquisition module 201 is used to obtain the clinical pathological data and laboratory analysis data of the patient, and to outline the target area on the L3 plane to obtain the skeletal muscle area, visceral fat area, subcutaneous fat area and intermuscular fat area on the L3 plane;

[0071] A clinical characteristic data processing module 203 is used to fill in missing data in clinical pathology data and laboratory analysis data based on a random forest chain equation multiple interpolation algorithm;

[0072] The univariate regression analysis screening module 205 is used to screen the variables using the univariate regression analysis technique to obtain key indicators for analyzing anastomotic leakage after colorectal cancer surgery; and to perform linear tests, multicollinearity tests and strong influence point analyses on the key indicators respectively;

[0073] A multi-factor regression analysis screening module 207 is used to perform a multi-factor regression analysis on key indicators with a P value less than a preset threshold based on a multi-factor regression analysis technique, and screen out independent predictive factors for predicting whether anastomotic leakage will occur in patients with colorectal cancer after surgery;

[0074] The anastomotic fistula occurrence probability prediction module 209 is used to construct and train an anastomotic fistula prediction model based on the screened independent prediction factors, and use the trained anastomotic fistula prediction model to predict the probability of anastomotic fistula occurring in colorectal cancer patients after surgery.

[0075] In order to facilitate understanding of the above technical solutions of the present invention, the above technical solutions of the present invention are further described below from the perspective of architecture and principle, as follows:

[0076] A method for predicting anastomotic fistula after colorectal cancer surgery is used to predict the probability of anastomotic fistula after colorectal cancer surgery, and has accurate and objective performance, comprising the following steps:

[0077] (1) Data from 3681 patients who underwent radical resection for colorectal cancer in our hospital from September 2013 to September 2021 were retrospectively collected, of which 60 patients were diagnosed with anastomotic leakage after surgery, referred to as the case group. A nested case-control study was then designed. According to the time and location of surgery, each patient with anastomotic leakage was randomly matched with two patients without anastomotic leakage as the control group. Inclusion criteria were: age > 18 years, primary colorectal adenocarcinoma confirmed by postoperative histopathological examination, abdominal CT examination within 2 weeks before surgery, and no neoadjuvant chemoradiotherapy before surgery. Exclusion criteria were: distant metastasis before surgery, missing clinical data, and palliative resection. Finally, a total of 180 patients were used to construct this prediction model.

[0078] (2) Clinical characteristics of patients with potential predictive value were included: Clinical pathological data included age, sex, BMI (body mass index), history of hypertension, history of coronary heart disease, history of smoking, history of drinking, NRS2002 score, tumor T stage, tumor N stage, tumor differentiation, surgical method, and surgical site. Laboratory data included preoperative neutrophil count, lymphocyte count, white blood cell count, platelet count, albumin level, alanine aminotransferase level, aspartate aminotransferase level, activated partial thromboplastin time, antithrombin III activity, fasting blood glucose level, and neutrophil-to-lymphocyte ratio. All of the above data were extracted from the electronic medical record system. The abdominal CT image files were transferred to medical image analysis tools (e.g., Tomovision's SliceOmatic software). Different CT value ranges were set when outlining different target areas. Specifically, when outlining the skeletal muscle area in the L3 plane, the CT value range is set to -29 to 150 HU; when outlining the visceral fat area in the L3 plane, the CT value range is set to -150 to -50 HU; when outlining the subcutaneous fat area and intermuscular fat area in the L3 plane, the CT value range is set to -190 to -30 HU. By using the semi-automatic outlining function to outline the target area in the L3 plane, the skeletal muscle area, visceral fat area, subcutaneous fat area, and intermuscular fat area in the L3 plane can be obtained.

[0079] (3) Pretreatment clinical characteristics: Data of some variables are missing, including white blood cell count, platelet count, albumin level, alanine aminotransferase level, and aspartate aminotransferase level. The random forest chain equation multiple interpolation algorithm is used to fill in the missing data, which is a method of multiple interpolation using a random forest model. Specifically, it is first necessary to load the "mice" and "randomforest" software packages in the statistical analysis software (such as R language software), and then divide the data into complete data without missing values ​​and missing data with missing values. The random forest model is trained using the complete data, and the missing data with missing values ​​are filled. The number of iterations is specified to be 5, so 5 interpolation data sets will be generated, and the missing values ​​in each data set will be slightly different. After multiple iterations, the stability of the model and the reliability of the interpolation results can be ensured. After the interpolation is completed, for each missing value, the average of the 5 predicted values ​​is taken as the data of the missing value for subsequent analysis. In general, 38 missing values ​​(0.78%) out of 4860 values ​​were estimated in this embodiment.

[0080] (4) Univariate regression analysis to screen relevant variables. Statistical analysis software (such as R language software) was used to process the data. Univariate regression analysis was performed on the data of the case group and the control group. The specific analysis results are shown in Table 1. The analysis results showed that 7 indicators were significantly correlated with the occurrence of anastomotic fistula after colorectal cancer surgery, and the P value was <0.05 (P value <0.05 means that the correlation between the indicator and the occurrence of anastomotic fistula is statistically significant). The 7 indicators are gender, white blood cell count, albumin level, fasting blood glucose level, NRS2002 score, skeletal muscle area at L3 plane, and visceral fat area at L3 plane.

[0081] Table 1 Comparison of the correlation test results between 28 indicators and the occurrence of anastomotic leakage in patients with colorectal cancer after surgery

[0082]

[0083]

[0084]

[0085]

[0086] P value, also known as significance probability value, is a statistic used to evaluate whether the independent variable has a significant effect on the dependent variable. In single-factor or multi-factor regression analysis, there are two hypotheses, one is the null hypothesis (H0), which is that the regression coefficient of the independent variable is equal to zero (that is, the independent variable has no significant effect on the dependent variable); the other is the alternative hypothesis (H1), which is that the regression coefficient of the independent variable is not equal to zero (that is, the independent variable has a significant effect on the dependent variable). The P value is calculated by the t-test in regression analysis. If the P value is less than the pre-set significance level (usually 0.05), the null hypothesis is rejected and it is believed that the independent variable has a significant effect on the dependent variable. If the P value is greater than the significance level, the null hypothesis is not rejected, and it is believed that there is insufficient evidence to show that the independent variable has a significant effect on the dependent variable, that is, the independent variable has no significant effect on the dependent variable.

[0087] (5) Before multivariate regression analysis, the above 7 indicators were subjected to linearity test, multicollinearity test and strong influence point analysis. For linearity test, the natural logarithm of the 5 continuous indicators was first calculated. The calculation formula was ln(x), where x was the above 5 continuous indicators. The calculation results were named ln_white blood cell count, ln_albumin level, ln_fasting blood glucose, ln_skeletal muscle area and ln_visceral fat area. The Box-Tidwell method was then used to test whether there was a linear relationship between the independent variable and the logistic transformation value of the dependent variable, that is, the interaction term between the continuous independent variable and its natural logarithm was included in the regression equation. In addition, the two categorical indicators were also included in the regression equation. In this embodiment, a total of 12 items were included in the model analysis, including five continuous independent variables: white blood cell count, albumin level, and fasting blood glucose level, categorical independent variables: gender and NRS2002 score, and five interaction terms: white blood cell count*ln_white blood cell count, albumin level*ln_albumin level, fasting blood glucose*ln_fasting blood glucose, skeletal muscle area*ln_skeletal muscle area, and visceral fat area*ln_visceral fat area. Therefore, in this embodiment, it is recommended that the significance level should be α=0.0042 (i.e., 0.05÷12). According to this significance level, the P values ​​of all interaction terms in this embodiment are higher than 0.0042, so there is a linear relationship between all continuous independent variables and the dependent variable. In addition, tolerance or variance inflation factor is used to diagnose multicollinearity between independent variables. If the tolerance is less than 0.1 or the variance inflation factor is greater than 10, it means that collinearity exists. In this embodiment, the tolerance of all variables is between 0.902 and 0.977, and the variance inflation factor is between 1.024 and 1.109, so there is no multicollinearity. Cook's distance is used to identify influential data points that may have a negative impact on the regression model. Any point whose Cook's distance exceeds 4 / n (where n is the total number of samples) is considered an outlier. The results of this embodiment show that the Cook's distance of each observation is less than the threshold, indicating that there are no strong influencing points in the data set (i.e., the set of various indicators collected from 180 patients).

[0088] There is a close logical relationship between the linear test, multicollinearity test, strong influence point analysis and subsequent multifactor logistic regression analysis to ensure the validity and reliability of the analysis results. The linear test is used to verify whether there is a linear relationship between the independent variable and the dependent variable. The multicollinearity test is used to check whether there is a high correlation (collinearity) between the independent variables. Strong influence point analysis is used to identify outliers or strong influence points that have a significant impact on the model. After the aforementioned tests ensure data quality, multifactor analysis is performed to build a robust and reliable multifactor logistic regression model to improve the model's explanatory power and prediction accuracy.

[0089] (6) Multivariate regression analysis. The seven indicators with a P value of less than 0.05 in the univariate regression analysis were further subjected to multivariate regression analysis to further screen the independent risk factors associated with predicting whether anastomotic leakage will occur in patients with colorectal cancer after surgery. The results showed that gender, fasting blood glucose, NRS2002 score, skeletal muscle area, and visceral fat area were independent risk factors for anastomotic leakage in patients with colorectal cancer after surgery. The results are shown in Table 2. Therefore, the above five indicators were fitted into a logistic regression model to establish a prediction model for predicting anastomotic leakage, and the prediction model was visualized as a nomogram (such as Figure 3 The nomogram showed that fasting blood glucose had the greatest impact on the outcome, followed by skeletal muscle area, visceral fat area, NRS2002 score and gender.

[0090] The steps of multivariate regression analysis are as follows:

[0091] Key indicators with P values ​​less than the preset threshold were included in the multivariate logistic regression model:

[0092] Logit(P(Y=1))=β0+β1X1+β2X2+···+β k X k

[0093] Among them, X1, X2, …X k is the key indicator initially screened; Y is the dependent variable, indicating whether anastomotic leakage occurs, Y = 1 indicates the occurrence of anastomotic leakage, P(Y = 1) indicates the probability that Y takes the value of 1, that is, the probability of anastomotic leakage; β0 is the intercept term, β1, β2, …β k The key indicators X1, X2, ...X are initially selected. k The regression coefficient of .

[0094] A stepwise regression method was used for variable selection, and variables with a P value less than 0.05 were selected and identified as independent predictors.

[0095] Table 2 Seven variables included in multivariate regression analysis

[0096]

[0097]

[0098] (7) Model validation. The Bootstrap method was used for internal validation, and the performance of the prediction model was tested by resampling 1000 times. Specifically, the core of the Bootstrap method is resampling with replacement. This can be easily implemented in R using the sample() function by simply adding the parameter replace = TRUE. Resampling with replacement was performed within the range of the original data and repeated 1000 times, and the mean value and 95% confidence interval of the area under the curve were calculated. According to the receiver operating characteristic (ROC) curve analysis, the area under the curve (AUC) of the nomogram was 0.848 (95% CI, 0.781-0.914), indicating that the nomogram can effectively predict the occurrence of anastomotic fistula (such as Figure 4 In addition, the calibration curve of the nomogram shows that in the model of this embodiment, the actual incidence of anastomotic leakage and the predicted incidence are highly consistent (as shown in FIG. Figure 5 The Hosmer-Lemeshow test P value was 0.343, indicating that there was no significant difference between the predicted value and the true value. Decision curve analysis (DCA) was then used to evaluate the application value of the prediction model in actual clinical decision-making. The results showed that the model can bring net benefits to patients in clinical practice and has good clinical application value.

[0099] (8) In this embodiment, five clinical indicators and imaging indicators that are significantly correlated with the outcomes are used to construct a prediction model for predicting the occurrence of anastomotic fistula in patients with colorectal cancer after surgery. The coefficients in the model equation are optimized to obtain a suitable model, which can more accurately predict the probability of anastomotic fistula in patients with colorectal cancer after surgery. The AUC value is as high as 0.848, which has good accuracy and reliability. It can facilitate clinicians to predict the probability of anastomotic fistula before surgery and guide treatment decisions.

[0100] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 6 As shown. The computer device includes a processor, a memory and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store static information and dynamic information data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the steps in the above method embodiment are implemented.

[0101] Those skilled in the art will understand that Figure 6The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0102] In addition, the present invention also provides a computer device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiment when executing the computer program.

[0103] In addition, the present invention further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps in the above method embodiment are implemented.

[0104] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0105] The present invention is not limited to the structures which have been described above and shown in the drawings, and various modifications and changes may be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A method for predicting anastomotic leakage after colorectal cancer surgery, characterized in that: include: Obtain the patient's clinical pathological data and laboratory analysis data, and outline the target area on the L3 plane to obtain the skeletal muscle area, visceral fat area, subcutaneous fat area, and intermuscular fat area on the L3 plane; Based on the random forest chain equation multiple imputation algorithm, missing data in clinical pathology data and laboratory analysis data were filled; The variables were screened by univariate regression analysis technology to obtain the key indicators for analyzing anastomotic leakage after colorectal cancer surgery; linear tests, multicollinearity tests and strong influence point analyses were performed on the key indicators; Based on multivariate regression analysis technology, multivariate regression analysis was performed on key indicators with significant probability values ​​less than the preset threshold, and independent predictive factors for predicting whether anastomotic leakage would occur in patients with colorectal cancer after surgery were screened; Based on the independent predictive factors screened, an anastomotic fistula prediction model was constructed and trained, and the trained anastomotic fistula prediction model was used to predict the probability of anastomotic fistula after surgery in patients with colorectal cancer.

2. The method for predicting anastomotic leakage after colorectal cancer surgery according to claim 1, characterized in that: The clinical pathological data include age, gender, body mass index, history of hypertension, history of coronary heart disease, history of smoking, history of drinking, NRS2002 score, tumor T stage, tumor N stage, tumor differentiation degree, surgical method and surgical site; The laboratory analysis data included preoperative neutrophil count, lymphocyte count, white blood cell count, platelet count, albumin level, alanine aminotransferase level, aspartate aminotransferase level, activated partial thromboplastin time, antithrombin III activity, fasting blood glucose level, and neutrophil-to-lymphocyte ratio.

3. The method for predicting anastomotic leakage after colorectal cancer surgery according to claim 1, characterized in that: When outlining the skeletal muscle area in the L3 plane, the CT value range was set to -29 to 150 HU; When outlining the visceral fat area in the L3 plane, the CT value range was set to -150 to -50 HU; When outlining the subcutaneous fat area and intermuscular fat area in the L3 plane, the CT value range was set to -190 to -30 HU.

4. The method for predicting anastomotic leakage after colorectal cancer surgery according to claim 1, characterized in that: The key indicators include gender, white blood cell count, albumin level, fasting blood glucose level, NRS2002 score, skeletal muscle area at L3 plane and visceral fat area at L3 plane; The independent predictive factors included gender, fasting blood glucose level, NRS2002 score, skeletal muscle area at L3 plane and visceral fat area at L3 plane.

5. The method for predicting anastomotic leakage after colorectal cancer surgery according to claim 1, characterized in that: The linear test, multicollinearity test and strong influence point analysis of the key indicators include: Based on the natural logarithm formula, the natural logarithm values ​​of the continuous indicators in the key indicators were calculated, and the Box-Tidwell method was used to test whether there was a linear relationship between the independent variables and the logistic transformation values ​​of the dependent variables; Use tolerance or variance inflation factors to diagnose multicollinearity among independent variables; Based on the comparison of each patient's Cook's distance with the preset distance threshold, we analyzed whether there were strong influencing points in the data set.

6. A prediction system for anastomotic leakage after colorectal cancer surgery, characterized in that: include: The clinical characteristic data acquisition module is used to obtain the patient's clinical pathological data and laboratory analysis data, and to outline the target area on the L3 plane to obtain the skeletal muscle area, visceral fat area, subcutaneous fat area and intermuscular fat area on the L3 plane; Clinical feature data processing module, used to fill in missing data in clinical pathology data and laboratory analysis data based on random forest chain equation multiple imputation algorithm; The univariate regression analysis screening module is used to screen variables using univariate regression analysis technology to obtain key indicators for analyzing anastomotic leakage after colorectal cancer surgery; and to perform linear tests, multicollinearity tests, and strong influence point analyses on the key indicators; A multi-factor regression analysis screening module is used to perform multi-factor regression analysis on key indicators with significance probability values ​​less than a preset threshold based on multi-factor regression analysis technology, and screen out independent predictive factors for predicting whether anastomotic fistula will occur in patients with colorectal cancer after surgery; The anastomotic fistula occurrence probability prediction module is used to construct and train an anastomotic fistula prediction model based on the independent predictive factors obtained by screening, and use the trained anastomotic fistula prediction model to predict the probability of anastomotic fistula occurring in colorectal cancer patients after surgery.

7. The prediction system for anastomotic leakage after colorectal cancer surgery according to claim 6, characterized in that: The clinical pathological data include age, gender, body mass index, history of hypertension, history of coronary heart disease, history of smoking, history of drinking, NRS2002 score, tumor T stage, tumor N stage, tumor differentiation degree, surgical method and surgical site; The laboratory analysis data included preoperative neutrophil count, lymphocyte count, white blood cell count, platelet count, albumin level, alanine aminotransferase level, aspartate aminotransferase level, activated partial thromboplastin time, antithrombin III activity, fasting blood glucose level, and neutrophil-to-lymphocyte ratio.

8. The prediction system for anastomotic leakage after colorectal cancer surgery according to claim 6, characterized in that: When the clinical characteristic data acquisition module outlines the target area in the L3 plane, the CT value range when outlining the skeletal muscle area in the L3 plane is set to -29 to 150HU; the CT value range when outlining the visceral fat area in the L3 plane is set to -150 to -50HU; and the CT value range when outlining the subcutaneous fat area and intermuscular fat area in the L3 plane is set to -190 to -30HU.

9. The prediction system for anastomotic leakage after colorectal cancer surgery according to claim 6, characterized in that: The key indicators include gender, white blood cell count, albumin level, fasting blood glucose level, NRS2002 score, skeletal muscle area at L3 plane and visceral fat area at L3 plane; The independent predictive factors included gender, fasting blood glucose level, NRS2002 score, skeletal muscle area at L3 plane and visceral fat area at L3 plane.

10. The prediction system for anastomotic leakage after colorectal cancer surgery according to claim 6, characterized in that: When performing linear tests, multicollinearity tests and strong influence point analyses on key indicators respectively, the univariate regression analysis screening module calculates the natural logarithm value of the continuous indicator in the key indicator based on the natural logarithm formula, and uses the Box-Tidwell method to test whether there is a linear relationship between the independent variable and the logistic transformation value of the dependent variable; uses tolerance or variance inflation factor to diagnose multicollinearity between independent variables; and analyzes whether there are strong influence points in the data set based on the comparison results of the Cook's distance of each patient with the preset distance threshold.

Images