Model for predicting colorectal cancer peritoneal metastasis by combining ctDNA and application thereof

Abstract

The invention discloses a prediction system, which is characterized in that the prediction system is a column diagram, the ctDNA state, CA125, tumor pathological N stage and nerve infiltration of a colorectal cancer patient are taken as prediction indexes, and the probability of peritoneal metastasis of the colorectal cancer patient is taken as a prediction result; a model for predicting colorectal cancer peritoneal metastasis in combination with ctDNA is provided on the basis of the prediction system, the model can accurately predict peritoneal metastasis of a colorectal cancer patient, the AUC value of the model reaches 0.784, the C-index value reaches 0.776, the Hosmer-Lemeshop test result is larger than 0.05, and the model can be used for predicting peritoneal metastasis of the colorectal cancer patient. The prediction of the model on peritoneal metastasis of the colorectal cancer patient has no obvious difference with the actual peritoneal metastasis of the colorectal cancer patient.

Description

Technical Field

[0001] The present invention relates to the field of biomedical technology, and in particular to a model for predicting peritoneal metastasis of colorectal cancer in combination with ctDNA and an application thereof. Background Art

[0002] Colorectal cancer is currently the third most common cancer in the world, with approximately 1.9 million new cases and 935,000 deaths reported in 2020. With advances in diagnostic technology and increased awareness of screening, it is possible to detect tumors earlier before they metastasize to distant sites. Among them, 4-19% of colorectal cancer patients will develop peritoneal metastasis.

[0003] Currently, colorectal cancer liver metastasis and lung metastasis can be followed up and evaluated using imaging methods such as ultrasound angiography, liver MRI, and chest and abdominal CT, and can be treated with surgical resection or CT scans to achieve a disease-free state. Compared with other tumor metastases, colorectal cancer peritoneal metastasis has certain limitations in imaging, and it is difficult to detect smaller peritoneal nodules in the early stages. In addition, patients with peritoneal metastasis often have a poor prognosis, with a median overall survival of only 6 to 9 months after diagnosis. The extent of peritoneal metastasis is negatively correlated with the survival of colorectal cancer patients. The greater the extent of peritoneal metastasis, the shorter the survival.

[0004] Circulating tumor DNA (ctDNA) has become a very promising potential prognostic biomarker due to its non-invasive, real-time and rapid characteristics. It can be used to evaluate the presence of micro-residual lesions, which are the main cause of cancer recurrence. Currently, the use of ctDNA to stratify the risk of patients receiving adjuvant therapy for colorectal cancer has shown significant advantages. Compared with the commonly used clinical pathological high-risk factors, ctDNA can more accurately predict tumor recurrence and can show abnormalities several months before the signs of recurrence are clearly shown in imaging examinations.

[0005] In breast cancer, a prediction model for adjuvant treatment of breast cancer has been constructed by integrating clinical pathological information and ctDNA gene mutations; in gastric cancer, there is also a prediction model for occult peritoneal metastasis of gastric cancer based on Lauren classification, CLDN18.2 score and serum CA125. As for colorectal cancer, there are currently peritoneal metastasis prediction models constructed by comprehensive imaging features, and there are also peritoneal metastasis prediction models constructed by comprehensive clinical pathological features, but both models require further imaging examinations to make preliminary predictions of peritoneal metastasis of colorectal cancer. And because of the limitations of peritoneal metastasis of colorectal cancer in imaging, indicators suitable for peritoneal metastasis of other cancers may not necessarily be suitable for the prediction of peritoneal metastasis of colorectal cancer.

[0006] Currently, imaging combined with serum oncology is commonly used in clinical practice to predict peritoneal metastasis of colorectal cancer. However, the increase in serum tumor markers also has a certain false positive rate and false negative rate, which also makes it impossible to accurately predict peritoneal metastasis of colorectal cancer.

[0007] Therefore, there is an urgent need for an accurate method that can predict long-term peritoneal metastasis of early colorectal cancer to provide a visual reference for clinical practice. Summary of the invention

[0008] The purpose of the present invention is to overcome the above-mentioned deficiencies of the prior art and to provide a model for predicting peritoneal metastasis of colorectal cancer in combination with ctDNA and its application.

[0009] A first object of the present invention is to provide a prediction system.

[0010] The second object of the present invention is to provide a model for predicting peritoneal metastasis of colorectal cancer by combining ctDNA.

[0011] The third object of the present invention is to provide the use of the above-mentioned prediction system and / or the above-mentioned model in the preparation of a product for predicting peritoneal metastasis of colorectal cancer.

[0012] A fourth object of the present invention is to provide a product.

[0013] A fifth object of the present invention is to provide a device.

[0014] A sixth object of the present invention is to provide an information data processing terminal.

[0015] In order to achieve the above object, the present invention is implemented by the following scheme:

[0016] A prediction system is a nomogram, which uses the ctDNA status, CA125, tumor pathological N stage and nerve infiltration of patients with stage I to III colorectal cancer as prediction indicators, and uses the probability of peritoneal metastasis in patients with colorectal cancer as a prediction result.

[0017] Among them, the classification of patients with stage I to III colorectal cancer refers to the expert consensus on clinical practice of ctDNA high-throughput sequencing (2022 edition).

[0018] Preferably, the ctDNA status of the patients with stage I to III colorectal cancer is the ctDNA status, CA125, pathological N stage and nerve infiltration of the patients with stage I to III colorectal cancer after radical surgery.

[0019] Preferably, the nomogram includes scales 1 to 7, scales 1 to 7 are parallel to each other, and each scale is a line segment and is engraved with scales;

[0020] The scale value of ruler 1 is 0-100, 0 and 100 are respectively set at the two ends of ruler 1, and the scale of ruler 1 is an equally divided scale; ruler 1 represents the score corresponding to the scale on ruler 2 to ruler 4;

[0021] Scale 2 indicates the ctDNA status, with scale values ​​of 0 and 1, 0 indicating ctDNA negative and 1 indicating ctDNA positive; when the scale value of scale 2 is 0, its score on scale 1 is 0; when the scale value of scale 2 is 1, its score on scale 1 is 100;

[0022] Scale 3 represents CA125, with scale values ​​of 0 and 1, 0 represents CA125 ≤ 35u / mL, and 1 represents CA125 > 35u / mL; the scale value of scale 3 is 0, and its score on scale 1 is 0; the scale value of scale 3 is 1, and its score on scale 1 is 54;

[0023] Scale 4 indicates the pathological N stage of the tumor, with scale values ​​of 0, 1 and 2, 0 indicating N0 stage, 1 indicating N1 stage, and 2 indicating N2 stage; the scale value of scale 4 is 0, and its score on scale 1 is 0; the scale value of scale 4 is 1, and its score on scale 1 is 41; the scale value of scale 4 is 2, and its score on scale 1 is 66;

[0024] Scale 5 represents nerve infiltration, and the scale values ​​are 0 and 1, 0 represents no nerve infiltration, and 1 represents the presence of nerve infiltration; the scale value of scale 5 is 0, and its score on scale 1 is 0; the scale value of scale 5 is 1, and its score on the scale is 35;

[0025] Scale 6 represents the total score, which is the sum of the scores corresponding to the scale values ​​on scales 2 to 5; the scale value of scale 6 is 0 to 280, and scale 6 is divided equally;

[0026] Scale 7 represents the probability of peritoneal metastasis of colorectal cancer, and the scale value is 0.1 to 0.8; the probability of peritoneal metastasis of colorectal cancer patients = the total score of scale 6 on scale 7 corresponding to the value.

[0027] More preferably, ctDNA positivity is defined as class I or class II mutations, with at least one somatic mutation detected; ctDNA negativity is defined as class III mutations or no somatic mutations detected in the ctDNA sample.

[0028] More preferably, the corresponding relationship between the probability of peritoneal metastasis of colorectal cancer shown in scale 7 and the total score of scale 6 is shown in the following table:

[0029] Total score of scale 6 0 35 41 54 66 76 89 95 Scale 7 Probability 0.03 0.07 0.08 0.1 0.12 0.14 0.18 0.19 Total score of scale 6 100 101 120 130 135 141 154 155 Scale 7 Probability 0.21 0.21 0.28 0.31 0.34 0.36 0.42 0.43 Total score of scale 6 166 176 195 201 220 230 255 Scale 7 Probability 0.48 0.53 0.61 0.64 0.72 0.75 0.83 .

[0030] The present invention also claims a model for predicting peritoneal metastasis of colorectal cancer in combination with ctDNA, comprising an acquisition module, a prediction module and a result output module;

[0031] The acquisition module is used to obtain clinical indicators of patients with stage I to III colorectal cancer; the clinical indicators shown include ctDNA status, CA125, tumor pathological N stage and nerve infiltration of colorectal cancer patients;

[0032] The prediction module is any of the prediction systems described above, and the clinical indicators acquired by the acquisition module are input into the prediction system as prediction indicators to obtain the probability of peritoneal metastasis in patients with colorectal cancer;

[0033] The result output module is used to output the probability of peritoneal metastasis of colorectal cancer patients obtained by the prediction module.

[0034] The present invention also seeks to protect the use of any of the above-mentioned prediction systems and / or the above-mentioned models in the preparation of a product for predicting peritoneal metastasis in patients with colorectal cancer.

[0035] Preferably, the peritoneal metastasis of colorectal cancer patients is peritoneal metastasis of stage I to III colorectal cancer patients after radical surgery.

[0036] The present invention also seeks to protect a product comprising any of the above-mentioned prediction systems and / or the above-mentioned models.

[0037] The present invention also claims protection for a device, comprising an acquisition unit, a storage unit and a processing unit, wherein the acquisition unit is used to acquire data of patients with colorectal cancer in stages I to III, wherein the data includes ctDNA status, CA125, tumor pathological N stage and nerve infiltration of the colorectal cancer patients;

[0038] The storage unit stores program instructions that can be executed by the processing unit;

[0039] The processing unit contains any of the above-mentioned prediction systems and / or the above-mentioned models;

[0040] When the program instructions are executed by the processing unit, the data of the colorectal cancer patient obtained by the acquisition unit is input into the processing unit to obtain the probability of peritoneal metastasis of the colorectal cancer patient.

[0041] The present invention also seeks to protect an information data processing terminal, comprising a memory and a processor, wherein the memory stores a computer program that can be executed on the processor;

[0042] When the computer program is executed by the processor, the processor executes any one of the prediction systems and / or models described above.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] The present invention provides a prediction system, which is a nomogram, using the ctDNA status, CA125, tumor pathological N stage and nerve infiltration of patients with stage I to III colorectal cancer as prediction indicators, and using the probability of peritoneal metastasis of patients with colorectal cancer as a prediction result; and based on the prediction system, a model for predicting peritoneal metastasis of colorectal cancer in combination with ctDNA is provided, and the model can achieve accurate prediction of peritoneal metastasis of patients with colorectal cancer, the AUC value of the model reaches 0.784, the C-index value reaches 0.776, and the Hosmer-Lemeshow test result is greater than 0.05, indicating that there is no significant difference between the model's prediction of peritoneal metastasis of patients with colorectal cancer and the actual peritoneal metastasis of patients with colorectal cancer. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 : The LASSO regression result and CLVASSO result diagram in Example 1; A is the LASSO regression result diagram; B is the CLVASSO result diagram;

[0046] Figure 2 The nomogram for predicting peritoneal metastasis in patients with colorectal cancer shown in Example 2;

[0047] Figure 3 is the ROC curve diagram in Example 3; A is the ROC curve diagram when peritoneal metastasis is predicted for colorectal cancer patients in the training set using the prediction system shown in Example 2; B is the ROC curve diagram when peritoneal metastasis is predicted for colorectal cancer patients in the validation set using the prediction system shown in Example 2; C is the ROC curve diagram when peritoneal metastasis is predicted by performing 500 internal samplings in the training set; D is the ROC curve diagram when peritoneal metastasis is predicted by performing 500 internal samplings in the validation set;

[0048] Figure 4 : is the calibration curve diagram in Example 3; A is the calibration curve when peritoneal metastasis is predicted for the training set; B is the calibration curve when peritoneal metastasis is predicted for the validation set;

[0049] Figure 5 Graphs of the DCA curve and the CIC curve in Example 3; A is the DCA curve for predicting peritoneal metastasis for the training set; B is the DCA curve for predicting peritoneal metastasis for the validation set; C is the CIC curve for predicting peritoneal metastasis for the training set; D is the CIC curve for predicting peritoneal metastasis for the validation set;

[0050] Figure 6: is the ROC curve diagram in comparative example 1; A is the ROC curve diagram when peritoneal metastasis is predicted for the training set in the experimental group and control groups 1 to 5; B is the ROC curve diagram when peritoneal metastasis is predicted for the validation set in the experimental group and control groups 1 to 5. DETAILED DESCRIPTION

[0051] The present invention is further described in detail below in conjunction with the accompanying drawings and specific examples of the specification. The examples are only used to explain the present invention and are not used to limit the scope of the present invention. The test methods used in the following examples are conventional methods unless otherwise specified; the materials and reagents used are reagents and materials that can be obtained from commercial channels unless otherwise specified.

[0052] The samples and corresponding information used in the embodiments of the present invention have been informed to the patients and authorized by the patients before use.

[0053] Example 1 Screening of Predictive Indicators for Peritoneal Metastasis in Colorectal Cancer Patients

[0054] 1. Experimental Methods

[0055] 1. Sample screening and indicator acquisition

[0056] Patients with colorectal cancer who were diagnosed with colorectal cancer and underwent radical surgery and ctDNA testing within 2 to 8 weeks after surgery in the Sixth Affiliated Hospital of Sun Yat-sen University between January 2010 and December 2022 were collected. A total of 299 eligible colorectal cancer patients were obtained as samples, and variable indicators of colorectal cancer patients in the sample set were collected, including clinical information, pathological information, and gene mutation information;

[0057] The clinical information included: age, gender, serum carcinoembryonic antigen (CEA), serum carbohydrate antigen 199 (CA199), serum carbohydrate antigen 125 (CA125), primary tumor site, adjuvant chemotherapy, recurrence, death, and peritoneal metastasis of colorectal cancer patients at the time of radical surgery;

[0058] Pathological information includes: pathological T stage after radical surgery, pathological N stage, pathological TNM stage, cancer nodules, vascular invasion information (whether vascular invasion occurs), and nerve invasion information (whether nerve invasion occurs);

[0059] Gene mutation information includes: plasma ctDNA results, mutation information of tissue TP53, tissue KRAS, tissue NRAS, tissue BRAF, tissue PIK3CA, tissue SMAD4, tissue FBXW7, plasma TP53, plasma KRAS, plasma NRAS, plasma BRAF, plasma PIK3CA and plasma FBXW7.

[0060] Gene mutation information was obtained through next-generation sequencing (NGS). Peritoneal metastasis was determined according to the response evaluation criteria in solid tumors (RECIST) version 1.1, in which the first occurrence of peritoneal metastasis was considered to be peritoneal metastasis regardless of whether it was combined with other organ metastases.

[0061] ctDNA positivity is defined as class I or class II mutations, with at least one somatic mutation detected; ctDNA negativity is defined as class III mutations or no somatic mutations detected in the ctDNA sample; class I mutations, class II mutations, and class III mutations are based on the clinical significance of tumor somatic mutations interpreted in the expert consensus on clinical practice of ctDNA high-throughput sequencing (2022 edition), and are graded according to companion diagnostics, clinical guidelines, databases or literature evidence. Tumor somatic gene mutations are divided into clear clinical significance (Class I), potential clinical significance (Class II), unclear clinical significance (Class III), and benign or possibly benign mutations (Class IV).

[0062] 2. Screening of predictive indicators for peritoneal metastasis in patients with colorectal cancer

[0063] SPSS25.0 and R-4.3.1 software were used to perform data processing, statistical analysis and result visualization on the variable indicators of the sample set described in step 1. The R software packages used were: survival, Caret, pROC, calibrate, MASS, rms, foreign, nricens, glmnet, Matrix, rmda, ggplot2, ResourceSelection, lattic and survex.

[0064] The sample set shown in step 1 was randomly divided into a training set (209 cases) and a validation set (90 cases) in a ratio of 7:3 using the Caret package. Based on the variable indicators (clinical information, pathological information, and gene mutation information) of each colorectal cancer patient, the chi-square test (x 2The differences between the variables in the training set and the validation set were analyzed by using the Fisher exact test or Fisher's exact test. If the sample frequency was <1, Fisher's exact test was used; if the sample frequency was between 1 and 5, the exact chi-square test was used; if the sample frequency was >5, the chi-square test was used to ensure that there was no statistically significant difference between the variables in the training set and the test set.

[0065] The Kolmogorov-Smirnov test was used to assess the normality of the variables in the sample set described in step 1. The variables with normal distribution were expressed as mean ± standard deviation (v ± s), and the t test was used for inter-group comparison. For continuous variables that did not meet the normality assumption in the variables, the median and interquartile range were used as descriptive statistics, and the Wilcoxon rank sum test was used as a nonparametric alternative for inter-group comparison. For categorical variables in the variables, counts and percentages were expressed based on the expected frequency, using the x test. 2 Test or Fisher's exact test was used for comparison between groups.

[0066] The training set was used as the test sample, and the glmnet package was used to perform multivariate logistic regression analysis and univariate logistic regression analysis on the variable indicators of colorectal cancer patients in the training set combined with the peritoneal metastasis of colorectal cancer patients. The variable indicators with statistically significant differences (p < 0.05) in the univariate logistic regression analysis were used as the initial screening variable indicators.

[0067] The primary screening variable indicators were included in the least absolute shrinkage and selection operator (LASSO) regression, and 10-fold cross-validation (CVLASSO) was performed. The minimum lambda was used as the criterion to determine the secondary screening variable indicators.

[0068] The multivariate stepwise regression analysis was performed on the secondary screening variables, and the minimum Akaike information criterion (AIC) was used to obtain the predictive indicators of peritoneal metastasis in patients with colorectal cancer.

[0069] 2. Experimental Results

[0070] The variable indicators in the sample set, training set and validation set as well as the chi-square test results are shown in Table 1.

[0071] Table 1 Variable indicators and chi-square test results

[0072]

[0073]

[0074] The results showed that there were no statistically significant differences in the variables of colorectal cancer patients in the training set and the validation set, and the P values ​​were all greater than 0.05.

[0075] The results of univariate logistic regression and multivariate logistic regression analysis of each variable index in the training set are shown in Table 2.

[0076] Table 2 Results of univariate logistic regression and multivariate logistic regression analysis

[0077]

[0078]

[0079]

[0080] According to the results in Table 2: In the univariate logistic regression analysis, ctDNA, CA125, pathological N stage, nerve infiltration, tissue TP53 mutation, plasma TP53 mutation, plasma KRAS mutation, plasma PIK3CA mutation, and plasma SMAD4 mutation were associated with peritoneal metastasis in patients with colorectal cancer, and the P values ​​were all less than 0.05.

[0081] The results of LASSO regression analysis are as follows: Figure 1 As shown in A, the CVLASSO results are Figure 1 As shown in B, the results show that the secondary screening variable indicators determined by the minimum lambda standard include: ctDNA, CA125, pathological stage N1, pathological stage N2, neural infiltration, tissue TP53 mutation, plasma TP53 mutation, plasma PIK3CA mutation, plasma SMAD4 mutation

[0082] A multivariate stepwise regression analysis was performed on the secondary screening variable indicators, and the predictive indicators of peritoneal metastasis in colorectal cancer patients obtained after incorporating factors into the screening model based on the minimum AIC principle included: ctDNA, CA125, pathological N stage (including pathological stage N0, pathological stage N1 and pathological stage N2) and neural infiltration of colorectal cancer patients.

[0083] Example 2 A prediction system for predicting peritoneal metastasis of colorectal cancer using ctDNA

[0084] 1. Construction of the prediction system

[0085] The prediction system for predicting peritoneal metastasis of colorectal cancer combined with ctDNA is a nomogram, as shown below:

[0086] Based on the prediction index of peritoneal metastasis of colorectal cancer patients obtained in Example 1, the rms package was used in combination with the multi-factor logistic regression analysis results shown in Example 1 to obtain the following: Figure 2The nomogram shown is used to predict peritoneal metastasis in patients with colorectal cancer; the corresponding scores of the prediction indicators of peritoneal metastasis in patients with colorectal cancer obtained in Example 1 in the nomogram are shown in Table 3.

[0087] Table 3 Corresponding scores of prediction indicators in the nomogram

[0088]

[0089]

[0090] Figure 2 The nomogram shown includes 7 scales, scales 1 to 7 are arranged in sequence and parallel to each other, and each scale is a line segment and is engraved with scales;

[0091] The scale value of ruler 1 is 0-100, 0 and 100 are respectively set at the two ends of ruler 1, and the scale of ruler 1 is an equally divided scale; ruler 1 represents the score corresponding to the scale on ruler 2 to ruler 4;

[0092] Scale 2 indicates the ctDNA status, with scale values ​​of 0 and 1, 0 indicating ctDNA negative and 1 indicating ctDNA positive; when the scale value of scale 2 is 0, its score on scale 1 is 0; when the scale value of scale 2 is 1, its score on scale 1 is 100;

[0093] Scale 3 represents CA125, with scale values ​​of 0 and 1, 0 represents CA125 ≤ 35u / mL, and 1 represents CA125 > 35u / mL; the scale value of scale 3 is 0, and its score on scale 1 is 0; the scale value of scale 3 is 1, and its score on scale 1 is 54;

[0094] Scale 4 represents the pathological N stage of the tumor, with scale values ​​of 0, 1 and 2, 0 represents pathological stage N0, 1 represents pathological stage N1, and 2 represents pathological stage N2; the scale value of scale 4 is 0, and its score on scale 1 is 0; the scale value of scale 4 is 1, and its score on scale 1 is 41; the scale value of scale 4 is 2, and its score on scale 1 is 66;

[0095] Scale 5 represents nerve infiltration, and the scale values ​​are 0 and 1, 0 represents no nerve infiltration, and 1 represents the presence of nerve infiltration; the scale value of scale 5 is 0, and its score on scale 1 is 0; the scale value of scale 5 is 1, and its score on the scale is 35;

[0096] Scale 6 represents the total score, which is the sum of the scores corresponding to the scale values ​​on scales 2 to 5; the scale value of scale 6 is 0 to 280, and scale 6 is divided equally;

[0097] Scale 7 represents the probability of peritoneal metastasis of colorectal cancer, and the scale value is 0.1 to 0.8; the probability of peritoneal metastasis of colorectal cancer patients = the total score of scale 6 on scale 7 corresponding to the value.

[0098] Scales 1 to 7 together constitute a nomogram for predicting peritoneal metastasis of colorectal cancer, that is, a prediction system for predicting peritoneal metastasis of colorectal cancer; the corresponding relationship between the total score shown in scale 6 and scale 7 is shown in Table 4.

[0099] Table 4 Correspondence between the total score shown in scale 6 and scale 7

[0100] Total score of scale 6 0 35 41 54 66 76 89 95 Scale 7 Probability 0.03 0.07 0.08 0.1 0.12 0.14 0.18 0.19 Total score of scale 6 100 101 120 130 135 141 154 155 Scale 7 Probability 0.21 0.21 0.28 0.31 0.34 0.36 0.42 0.43 Total score of scale 6 166 176 195 201 220 230 255 Scale 7 Probability 0.48 0.53 0.61 0.64 0.72 0.75 0.83

[0101] 2. How to use the prediction system

[0102] The ctDNA, CA125, pathological N stage and neural infiltration of colorectal cancer patients were obtained as predictive indicators, and combined with the prediction system (nomogram) constructed in step 1 to obtain the probability of peritoneal metastasis in colorectal cancer patients.

[0103] Example 3 Application of a prediction system combining ctDNA to predict peritoneal metastasis of colorectal cancer in predicting peritoneal metastasis

[0104] 1. Experimental Methods

[0105] The training set (209 colorectal cancer patients) and validation set (90 colorectal cancer patients) shown in Example 1 were used as test objects, respectively, to obtain the predictive indicators of each colorectal cancer patient in the test set and the validation set, including ctDNA, CA125, pathological N stage and neural infiltration of colorectal cancer patients; and the actual peritoneal metastasis of each colorectal cancer patient in the training set and the validation set was recorded respectively.

[0106] The prediction indicators of each colorectal cancer patient in the training set and the validation set are respectively used, and the prediction system for predicting colorectal cancer peritoneal metastasis constructed in Example 2 is combined with the usage method shown in step 2 of Example 2 to obtain the peritoneal metastasis probability of each colorectal cancer patient in the training set and the validation set.

[0107] The pROC package was used to draw the ROC curves of the training set and the validation set, combining the peritoneal metastasis model prediction of each colorectal cancer patient in the training set and the validation set with the actual peritoneal metastasis of the corresponding colorectal cancer patients.

[0108] When the ROC curve is drawn for the training set, the cutoff value is 0.091; when the ROC curve is drawn for the test set, the cutoff value is 0.117.

[0109] According to the above method, 500 internal samples were taken from colorectal cancer patients in the training set and validation set to predict peritoneal metastasis, and a ROC curve was drawn based on the actual peritoneal metastasis of colorectal cancer patients.

[0110] The rms package was used to draw the calibration curves of the training set and the validation set, combining the peritoneal metastasis model predictions of each colorectal cancer patient in the training set and the validation set with the actual peritoneal metastasis of the corresponding colorectal cancer patients.

[0111] The Hosmer-Lemeshow test was performed using the ResourceSelection package based on the difference between the model's predicted peritoneal metastasis and the actual peritoneal metastasis of the corresponding colorectal cancer patients.

[0112] Combined with the prediction of peritoneal metastasis model of each colorectal cancer patient in the training set and validation set and the actual peritoneal metastasis of the corresponding colorectal cancer patients, the rmda package was used to draw the DCA curve and CIC curve of the training set and validation set respectively.

[0113] 2. Experimental Results

[0114] When the prediction system shown in Example 2 is used to predict peritoneal metastasis in patients with colorectal cancer in the training set, the specificity is 0.598 and the sensitivity is 0.886. The ROC curve is shown in FIG. Figure 3 As shown in A, the results show that the area under the curve (AUC) is 0.808; when the prediction system shown in Example 2 is used to predict peritoneal metastasis in colorectal cancer patients in the validation set, the specificity is 0.608 and the sensitivity is 0.875. The ROC curve is shown in Figure 3 As shown in B, the results showed that the area under the curve (AUC) was 0.784.

[0115] The ROC curve for peritoneal metastasis prediction using 500 internal samplings in the training set is shown below: Figure 3 As shown in C, the ROC curve of peritoneal metastasis prediction was performed 500 times in the validation set. Figure 3 As shown in D, the results show that the areas under the curve are 0.809 and 0.776 respectively.

[0116] The calibration curve for peritoneal metastasis prediction based on the training set is as follows Figure 4 As shown in A, the calibration curve for peritoneal metastasis prediction for the validation set is as follows Figure 4 B. The results show that the model-predicted probability of peritoneal metastasis is basically consistent with the actual probability of peritoneal metastasis, indicating that the prediction system for predicting peritoneal metastasis of colorectal cancer in combination with ctDNA as shown in Example 2 can accurately predict peritoneal metastasis in patients with colorectal cancer.

[0117] The Hosmer-Lemeshow test result for predicting peritoneal metastasis for the training set was 0.394, and the Hosmer-Lemeshow test result for predicting peritoneal metastasis for the validation set was 0.261, both greater than 0.05, indicating that the prediction system for predicting peritoneal metastasis of colorectal cancer combined with ctDNA shown in Example 2 has a good goodness of fit and no significant difference with the actual observed values.

[0118] The DCA curve for peritoneal metastasis prediction for the training set is as follows Figure 5 As shown in A, the DCA curve for peritoneal metastasis prediction for the validation set is as follows Figure 5 As shown in B, the CIC curve for peritoneal metastasis prediction for the training set is as follows Figure 5 As shown in C, the CIC curve for peritoneal metastasis prediction for the validation set is as follows Figure 5 As shown in D.

[0119] The results showed that when the model prediction probability threshold was between 10% and 90%, the prediction system for predicting colorectal cancer peritoneal metastasis combined with ctDNA shown in Example 2 had the greatest benefit and had a high diagnostic accuracy within this range; this indicated that the prediction system for predicting colorectal cancer peritoneal metastasis combined with ctDNA shown in Example 2 had a high diagnostic accuracy and was reliable.

[0120] Comparative Example 1: Prediction of peritoneal metastasis in patients with colorectal cancer by combining different variable indicators

[0121] 1. Experimental Methods

[0122] The training set (209 colorectal cancer patients) and the validation set (90 colorectal cancer patients) shown in Example 1 were used as test objects, and nomograms were constructed by combining different combinations of variable indicators to predict the probability of peritoneal metastasis in colorectal cancer patients, as follows:

[0123] Experimental group: The prediction system for predicting peritoneal metastasis of colorectal cancer constructed in Example 2 was used in combination with the method shown in step 2 of Example 2 to predict the probability of peritoneal metastasis of colorectal cancer patients in the training set and the validation set, respectively. The ROC curve was drawn based on the actual peritoneal metastasis of each colorectal cancer patient.

[0124] Control group 1: According to the construction method described in step 1 of Example 2, the predictive indicators of peritoneal metastasis of colorectal cancer patients were modified to CA125 and neural infiltration of colorectal cancer patients, and a prediction system 1 was constructed to predict the probability of peritoneal metastasis of colorectal cancer patients in the training set and the validation set, respectively, and the ROC curve was drawn based on the actual peritoneal metastasis of each colorectal cancer patient.

[0125] Control group 2: According to the construction method described in step 1 of Example 2, the predictive indicators of peritoneal metastasis of colorectal cancer patients were modified to CA125, pathological N stage and neural infiltration of colorectal cancer patients, and a prediction system 2 was constructed to predict the probability of peritoneal metastasis of colorectal cancer patients in the training set and the validation set, respectively, and the ROC curve was drawn based on the actual peritoneal metastasis of each colorectal cancer patient.

[0126] Control group 3: According to the construction method described in step 1 of Example 2, the predictive indicators of peritoneal metastasis of colorectal cancer patients were modified to CA125, pathological N stage, cancer nodule information (yes / no) and nerve infiltration of colorectal cancer patients, and a prediction system 3 was constructed to predict the probability of peritoneal metastasis of colorectal cancer patients in the training set and the validation set, respectively, and the ROC curve was drawn based on the actual peritoneal metastasis of each colorectal cancer patient.

[0127] Control group 4: A model for predicting peritoneal metastasis based on TNM staging of colorectal cancer patients was constructed using logistic regression. The probability of peritoneal metastasis of colorectal cancer patients in the training set and validation set was predicted respectively, and the ROC curve was drawn based on the actual peritoneal metastasis of each colorectal cancer patient.

[0128] Control group 5: According to the construction method described in step 1 of Example 2, the predictive indicators of peritoneal metastasis of colorectal cancer patients were modified to ctDNA, pathological N stage, and neural infiltration of colorectal cancer patients, and a prediction system 4 was constructed to predict the probability of peritoneal metastasis of colorectal cancer patients in the training set and the validation set, respectively, and the ROC curve was drawn based on the actual peritoneal metastasis of each colorectal cancer patient.

[0129] 2. Experimental Results

[0130] The ROC curves for the prediction of peritoneal metastasis in the experimental group and control group 1 to control group 5 for the training set are shown in Figure 2. Figure 6 As shown in A, the ROC curves for peritoneal metastasis prediction in the experimental group and control group 1 to control group 5 for the validation set are as follows Figure 6 As shown in B.

[0131] The results show that the prediction system for predicting peritoneal metastasis of colorectal cancer shown in Example 2 of the present invention can achieve more accurate prediction of peritoneal metastasis of colorectal cancer patients. Only by combining the ctDNA, CA125, pathological N stage and neural infiltration of colorectal cancer patients can accurate prediction of peritoneal metastasis be achieved; changes in indicator variables will cause a significant decrease in the prediction results.

[0132] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit the protection scope of the present invention. For ordinary technicians in this field, other different forms of changes or modifications can be made based on the above descriptions and ideas. It is not necessary and impossible to list all the implementation methods here. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A prediction system, characterized in that: The prediction system is a nomogram, which uses the ctDNA status, CA125, tumor pathological N stage and neural infiltration of patients with stage I to III colorectal cancer as prediction indicators, and the probability of peritoneal metastasis in patients with colorectal cancer as the prediction result.

2. The prediction system according to claim 1, characterized in that The ctDNA status of the patients with stage I to III colorectal cancer refers to the ctDNA status, CA125, pathological N stage and nerve infiltration of the patients with stage I to III colorectal cancer after radical surgery.

3. The prediction system according to claim 1, characterized in that The nomogram includes scales 1 to 7, scales 1 to 7 are parallel to each other, and each scale is a line segment and is engraved with scales; The scale value of ruler 1 is 0-100, 0 and 100 are respectively set at the two ends of ruler 1, and the scale of ruler 1 is an equally divided scale; ruler 1 represents the score corresponding to the scale on ruler 2 to ruler 4; Scale 2 indicates the ctDNA status, with scale values ​​of 0 and 1, 0 indicating ctDNA negative and 1 indicating ctDNA positive; when the scale value of scale 2 is 0, its score on scale 1 is 0; when the scale value of scale 2 is 1, its score on scale 1 is 100; Scale 3 represents CA125, with scale values ​​of 0 and 1, 0 represents CA125 ≤ 35u / mL, and 1 represents CA125 > 35u / mL; the scale value of scale 3 is 0, and its score on scale 1 is 0; the scale value of scale 3 is 1, and its score on scale 1 is 54; Scale 4 indicates the pathological N stage of the tumor, with scale values ​​of 0, 1 and 2, 0 indicating N0 stage, 1 indicating N1 stage, and 2 indicating N2 stage; the scale value of scale 4 is 0, and its score on scale 1 is 0; the scale value of scale 4 is 1, and its score on scale 1 is 41; the scale value of scale 4 is 2, and its score on scale 1 is 66; Scale 5 represents nerve infiltration, and the scale values ​​are 0 and 1, 0 represents no nerve infiltration, and 1 represents the presence of nerve infiltration; the scale value of scale 5 is 0, and its score on scale 1 is 0; the scale value of scale 5 is 1, and its score on the scale is 35; Scale 6 represents the total score, which is the sum of the scores corresponding to the scale values ​​on scales 2 to 5; the scale value of scale 6 is 0 to 280, and scale 6 is divided equally; Scale 7 represents the probability of peritoneal metastasis of colorectal cancer, and the scale value is 0.1 to 0.8; the probability of peritoneal metastasis of colorectal cancer patients = the total score of scale 6 on scale 7 corresponding to the value.

4. The prediction system according to claim 3, characterized in that The corresponding relationship between the probability of peritoneal metastasis of colorectal cancer shown in scale 7 and the total score of scale 6 is shown in the following table: 。 5. A model for predicting peritoneal metastasis of colorectal cancer combined with ctDNA, characterized in that: It includes an acquisition module, a prediction module and a result output module; The acquisition module is used to acquire clinical indicators of patients with colorectal cancer; the clinical indicators shown include ctDNA status, CA125, tumor pathological N stage and nerve infiltration of patients with stage I to III colorectal cancer; The prediction module is the prediction system according to any one of claims 1 to 4, wherein the clinical index acquired by the acquisition module is input into the prediction system as a prediction index to obtain the probability of peritoneal metastasis in patients with colorectal cancer; The result output module is used to output the probability of peritoneal metastasis of colorectal cancer patients obtained by the prediction module.

6. Use of the prediction system according to any one of claims 1 to 4 and / or the model according to claim 5 in the preparation of a product for predicting peritoneal metastasis in patients with colorectal cancer.

7. The use according to claim 6, characterized in that: The peritoneal metastasis of colorectal cancer patients is peritoneal metastasis of patients with stage I to III colorectal cancer after radical surgery.

8. A product, characterized in that Contains the prediction system according to any one of claims 1 to 4 and / or the model according to claim 5.

9. A device comprising an acquisition unit, a storage unit and a processing unit, characterized in that: The acquisition unit is used to acquire data of colorectal cancer patients, wherein the data includes ctDNA status, CA125, tumor pathological N stage and nerve infiltration of patients with stage I to III colorectal cancer; The storage unit stores program instructions that can be executed by the processing unit; The processing unit comprises the prediction system according to any one of claims 1 to 4 and / or the model according to claim 5; When the program instructions are executed by the processing unit, the data of the colorectal cancer patient obtained by the acquisition unit is input into the processing unit to obtain the probability of peritoneal metastasis of the colorectal cancer patient.

10. An information data processing terminal, comprising a memory and a processor, characterized in that: The memory stores a computer program executable on the processor; When the computer program is executed by the processor, the processor executes the prediction system according to any one of claims 1 to 4 and / or the model according to claim 5.

Images