Blood metabolism markers and application thereof

Abstract

The invention relates to a group of blood metabolism markers and application thereof. The blood metabolism markers comprise two or more than two metabolism markers of lauric acid, phosphatidylcholine PC (14: 1 (9Z) / 24: 1 (15Z)), phosphatidylcholine PC 34: 2, phosphatidylcholine PC 35: 2 and cis-8, 11, 14, 17-eicosatetraenoic acid. According to the invention, it is found that the relative expression level of metabolites in blood is closely related to capecitabine-related hand-foot syndrome, a metabolite marker combination is screened by a method of taking a common intersection through machine learning, a group of more refined markers are obtained, and a robust hand-foot syndrome adverse reaction prediction model is constructed in combination with a classical Logistic method.

Description

Technical Field

[0001] The present invention belongs to the field of colorectal cancer diagnosis, and particularly relates to a group of blood metabolic markers and applications thereof. Background Art

[0002] Colorectal cancer (CRC) is one of the most common malignancies worldwide, with the third highest incidence and second highest mortality rate among cancers worldwide (Freddie, Bray, Jacques, et al. Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. [J]. CA: a cancer journal for clinicians, 2018). Postoperative adjuvant chemotherapy is an important cornerstone of CRC treatment. The National Comprehensive Cancer Network (NCCN) guidelines recommend capecitabine as a first-line treatment for advanced colorectal cancer. Postoperative adjuvant chemotherapy plays an important role in reducing the risk of CRC recurrence and mortality. However, it also brings a number of adverse reactions. Hand-foot syndrome (HFS) is the most common adverse reaction to capecitabine chemotherapy, with an incidence rate as high as 70% (Li M, Sun X, Yao H, et al. Genomic methylation variations predict the susceptibility of six chemotherapy-related adverse effects and cancer development for Chinese colorectal cancer patients[J]. Toxicology and Applied Pharmacology, 2021, 427(Pt2):115657.), ranking among the top three adverse reactions associated with capecitabine chemotherapy. The occurrence of HFS not only places a huge burden on clinicians, but also affects patients' quality of life, reduces their medication compliance, and affects the efficacy of chemotherapy drugs and patient prognosis.

[0003] Currently, clinical methods for predicting adverse reactions mainly focus on genetic polymorphisms and TDM testing. Genetic polymorphism biomarkers such as DPD (rs3918290) are mainly targeted at European and American populations and are not applicable to the Chinese population (Ezzeldin, HHMethylation of the DPYD promoter: an alternative mechanism for dihydropyrimidine dehydrogenase deficiency in cancer patients. [J]. Clinical Cancer Research An Official Journal of the American Association for Cancer Research, 2005, 11(24): 8699.). TDM testing has the limitations of complex operation and delayed results.

[0004] Metabolomics, a crucial component of systems biology, can reveal overall metabolic differences in biological systems that are not readily apparent at the genetic level by measuring the levels of low-molecular-weight metabolites in body fluids such as blood and urine. Metabolomics has become an important technical tool in areas such as early cancer diagnosis, drug safety evaluation, and personalized toxicology research. Currently, many potential biomarker prediction models, including metabolomics-based blood, urine, feces, and tissue samples, have been developed for the early diagnosis of CRC.

[0005] However, there are no biomarkers or prediction models for the adverse effects of capecitabine-related hand-foot syndrome in CRC. Therefore, while ensuring the effectiveness of chemotherapy, it is urgent to find potential biomarkers with high sensitivity and specificity to effectively prevent adverse reactions to chemotherapy. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a group of blood metabolic markers and their applications. The relative expression levels of this group of metabolic markers are closely related to capecitabine-associated hand-foot syndrome. The metabolite marker combinations are screened by taking the common intersection method through machine learning to obtain a more refined group of multiple markers. At the same time, a robust hand-foot syndrome adverse reaction prediction model is constructed in combination with the classic logistic method.

[0007] The present invention provides a group of blood metabolic markers, including two or more metabolic markers of dodecanoic acid, phosphatidylcholine PC (14:1(9Z) / 24:1(15Z)), phosphatidylcholine PC 34:2, phosphatidylcholine PC 35:2 and cis-8,11,14,17-eicosatetraenoic acid.

[0008] Preferably, three or more metabolic markers are included among dodecanoic acid, phosphatidylcholine PC (14:1(9Z) / 24:1(15Z)), phosphatidylcholine PC 34:2, phosphatidylcholine PC 35:2 and cis-8,11,14,17-eicosatetraenoic acid.

[0009] More preferred ones include dodecanoic acid, phosphatidylcholine PC (14:1(9Z) / 24:1(15Z)), phosphatidylcholine PC 34:2, phosphatidylcholine PC 35:2 and cis-8,11,14,17-eicosatetraenoic acid.

[0010] Furthermore, a multi-factor logistic regression model is constructed based on the metabolic markers to output prediction results.

[0011] Furthermore, the model is:

[0012] Logit[P=HFS]=3.452+15.489*[dodecanoic acid]+8.947*[PC(14:1(9Z) / 24:1(15Z))]+2.674*[PC 34:2]+8.143*[PC 35:2]-9.803*[cis-8,11,14,17-eicosatetraenoic acid], where [P=HFS] represents the predictive efficacy of hand-foot syndrome based on the above biomarker combination, with a distribution interval of 0 to 1, and the cutoff value of [P=HFS] is set to 0.899; [dodecanoic acid], [PC(14:1(9Z) / 24:1(15Z))], [PC 34:2], [PC 35:2], and [cis-8,11,14,17-eicosatetraenoic acid] respectively represent the relative contents of the above metabolite markers in the blood.

[0013] Furthermore, when [P=HFS]≥0.899, it indicates that the person being tested has a risk of developing an adverse reaction of hand-foot syndrome; when [P=HFS]<0.899, it indicates that the person being tested does not have a risk of developing an adverse reaction of hand-foot syndrome.

[0014] The present invention also provides the use of a group of blood metabolic markers in the preparation of a kit for predicting the adverse reaction of hand-foot syndrome in patients with rectal cancer who receive postoperative capecitabine-adjuvant chemotherapy.

[0015] The present invention applies non-targeted metabolomics technology, screens markers through LASSO and RF, and constructs a prediction model for capecitabine-related hand-foot syndrome adverse reactions based on multivariate logistic regression. Clinical data Y and metabolomics preoperative blood data X of colorectal cancer patients receiving postoperative capecitabine-based adjuvant chemotherapy were obtained, and the data of the embodiment were preprocessed to obtain the Entire Set data set that can be used for subsequent analysis; the samples in the Entire Set were randomly stratified and split to generate a Training Set (80%) and a Validation Set (20%); in the Entire Set, potential biomarkers of hand-foot syndrome adverse reactions were preliminarily identified (differential analysis, PCA, etc.); potential markers were first screened using univariate logistic regression in the Training Set; then, potential markers were secondarily screened using the leastabsolute shrinkage and selection operator (LASSO) and random forest (RF) machine learning algorithms to avoid feature collinearity generated by high-dimensional data; further, the intersection of markers obtained by the LASSO and RF algorithms was selected as the final predictive marker for model construction, a multivariate logistic regression model was constructed, and the prediction results were output and verified in the Validation Set to provide a reference basis for clinical implementation.

[0016] Beneficial effects

[0017] 1. The blood samples used in the present invention are easy to obtain and cause less trauma to the tested population. In addition, compared with urine and fecal samples, the present invention has the advantages of being relatively more acceptable to the tested population and having higher compliance, and is also more easily accepted by testing staff. It is very suitable for large-scale population screening for adverse reactions of hand-foot syndrome associated with capecitabine chemotherapy after colorectal cancer surgery.

[0018] 2. Compared with the existing adverse reaction marker screening methods, the present invention first eliminates the influence of various potential confounding factors, and can better reflect the relevant effects of postoperative capecitabine-based adjuvant chemotherapy on adverse reactions, thereby reducing the false positive rate and false negative rate.

[0019] 3. The present invention uses preoperative blood samples of colorectal cancer patients as research subjects. Preoperative blood samples can better reflect the differences in individual physical functions.

[0020] 4. This study, for the first time, discovered that the relative expression levels of metabolites in the blood are closely associated with capecitabine-associated hand-foot syndrome. Using machine learning and common intersection methods to screen metabolite marker combinations, a more refined combination of five markers was obtained. Combined with the classic logistic method, a robust prediction model for adverse reactions to hand-foot syndrome was constructed, and its predictive efficacy was verified.

[0021] 5. When the cutoff value of [P=HFS] in the metabolic marker combination prediction model provided by the present invention is 0.899, the prediction model has the best predictive efficiency. Using multiple indicators for model evaluation (also applicable to data sets with unbalanced grouping), the area under the ROC curve reached 0.991 (95% CI: 0.970-1.000) and 0.848 (95% CI: 0.652-1.000) in the training set and validation set, respectively. The accuracy, precision, and recall rates all reached above 0.8, showing good prediction effect and accurate and reliable results. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 The figure is a flow chart of the method for using the markers of the present invention.

[0023] Figure 2 This is a PCA diagram of blood differential metabolites in the Entire Set in an embodiment of the present invention.

[0024] Figure 3 3. ROC curves of the capecitabine-related hand-foot syndrome adverse reaction marker prediction model in the Training Set, Validation Set and Entire Set in the embodiment of the present invention. DETAILED DESCRIPTION

[0025] Below in conjunction with specific embodiment, further set forth the present invention.Should be understood that these embodiments are only used to illustrate the present invention and are not used in limiting the scope of the present invention.In addition, should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall equally within the scope limited by the appended claims of the application.

[0026] Example 1

[0027] Construction of a prediction model for hand-foot syndrome in patients undergoing capecitabine chemotherapy after colorectal cancer surgery. Figure 1 As shown:

[0028] Step 1: Dataset Construction. Clinical data (Y) (including each patient's graded record of adverse reactions to hand-foot syndrome and general condition) and preoperative blood data (X) of non-targeted metabolomics were obtained from patients with colorectal cancer who received postoperative capecitabine-based adjuvant chemotherapy. Clinical data (complete follow-up records of adverse reactions to hand-foot syndrome and general condition (age, height, weight, BMI, tumor location, and tumor stage) and preoperative blood data (X) of 63 patients with colorectal cancer who underwent radical surgery and received postoperative capecitabine-based adjuvant chemotherapy between January 2018 and April 2019) were collected.

[0029] Patients' clinical data: According to the patients' adverse reaction records during chemotherapy, a total of 33 cases had grade 1 hand-foot syndrome adverse reactions and 9 cases had grade 2-3 hand-foot syndrome adverse reactions.

[0030] Analysis of preoperative blood samples: 15 mL of fasting blood was collected from patients within 1-3 days before surgery in EDTA tubes and stored at -80°C for subsequent analysis.

[0031] Before the experiment, thaw blood samples. Pipette 100 μL of plasma sample and add 400 μL of methanol (containing the internal standard: choline-d4, 2.0 μg / mL). Mix and vortex for 60 seconds. Centrifuge at 14,000 g for 15 minutes at 4°C. Finally, collect the supernatant and store at −20°C for subsequent LC-MS / MS analysis. QC samples were prepared by mixing an equal volume (10 μL) of each study sample.

[0032] LC-MS / MS detection conditions were as follows: UHPLC-Q-Exactive-MS technology was used, with an Agilent 1290 UHPLC system (Agilent Technologies, Inc., USA) and a Q Exactive mass spectrometer (Thermo Fisher Scientific, USA). Columns: ACQUITY UPLC BEH C8 column (1.7 μm, 2.1 × 100 mm, Waters, USA) was used in positive ion mode, and ACQUITY HSS T3 column (1.8 μm, 2.1 × 100 mm, Waters, USA) was used in negative ion mode. Column temperature: 50°C; flow rate: 0.3 mL / min; injection volume: 10 μL. Mobile phases: in positive ion mode: phase A: water (containing 0.1% formic acid), phase B: acetonitrile (containing 0.1% formic acid); in negative ion mode: phase A: water (containing 6.5 mM ammonium carbonate), phase B: 95% methanol (containing 6.5 mM ammonium carbonate). A gradient elution method was used with the following settings: in positive ion mode, 0-1 min (5% B), 1-24 min (5-100% B), 24.1-27.5 min (100% B), 27.6-30 min (100-5% B); in negative ion mode, 0-1 min (5% B), 1-18 min (5-100% B), 18.1-22 min (100% B), 22.1-25 min (100-5% B). Each injection volume was 10 μL, and all samples were injected in a random order. To reduce systematic error and ensure system stability, a QC sample was run between every 10 samples during actual sample testing, and an elution step was added after every 10 sample runs.

[0033] Step 2: Data preprocessing. The metabolomics data in the dataset were subjected to internal standard correction (choline-d4 as the internal standard), 80% rule processing (retaining characteristic ion peaks with a frequency greater than 80% in any group), quality control (eliminating metabolic features with a coefficient of variation CV>0.4 in QC samples), normalization (log10 transformation, Pareto scaling), compound identification, data outlier detection (IQR method to detect outliers within each group) and filling (KNN-sample wise), and confounder correction (using the removeBatchEffect function in the limma package to eliminate the effects of the following confounders: age, height, weight, BMI, tumor site, and tumor stage). This yielded the initial exploratory dataset, the Entire Set. The Entire Set was then randomly stratified and split into a training set (80%) and a validation set (20%).

[0034] Step 3: Identification of differential metabolites related to adverse reactions of hand-foot syndrome. First, a differential analysis was performed in the Entire Set to analyze the differences in blood metabolite levels between the two groups with hand-foot syndrome (grade 1, grade 2-3) and the group without hand-foot syndrome (grade 0), as well as between the grade 1 hand-foot syndrome group and the grade 2-3 hand-foot syndrome group. The results of the three pairwise comparison group differential analysis were integrated to screen differential metabolites related to hand-foot syndrome, that is, at least in any pairwise comparison of any group (grade 1 hand-foot syndrome vs grade 0, grade 2-3 hand-foot syndrome vs grade 0, grade 2-3 hand-foot syndrome vs grade 1), the following thresholds were met: P < 0.05 and |log10FC| > 0 (the limma package was used for differential analysis here). Then, the exogenous metabolites were eliminated. Principal component analysis (PCA) was used to analyze the inter-group differences in each group of hand-foot syndrome, and the results showed that ( Figure 2 ), there were significant differences in the relative contents of endogenous metabolites between the hand-foot syndrome group (grade 1, 2-3) and the hand-foot syndrome-free group (grade 0).

[0035] Step 4: Preliminary screening of potential markers related to adverse reactions of hand-foot syndrome. In the Training Set, the differential metabolites obtained from the above screening were subjected to univariate logistic regression analysis to screen out potential markers related to adverse reactions of hand-foot syndrome (P < 0.05).

[0036] Step 5: Finalization of candidate potential biomarkers related to adverse reactions of hand-foot syndrome. The metabolites obtained in Step 4 were further screened using two machine learning algorithms, LASSO and RF, to identify the top 10 most important potential biomarkers for each algorithm.

[0037] Among them, the LASSO algorithm is based on the "non-zero regression coefficient" through 10-fold cross validation; the random forest generates 1000 trees and selects the top N most important candidate markers of each algorithm based on the "Mean Decrease Accuracy" indicator, where the value of N is 10.

[0038] To further determine the optimal marker combination, the intersection of the top 10 important features of LASSO and RF was selected for a second screening of candidate markers, and finally a total of 5 candidate predictive markers were obtained (Table 1).

[0039] Table 1 Potential predictive markers for adverse reactions to hand-foot syndrome

[0040]

[0041] The results showed that the relative contents of the above five predictive markers had significant differences and important biological significance between the group with hand-foot syndrome (grade 1, grade 2-3) and the group without hand-foot syndrome (grade 0).

[0042] Step 6: Construction and evaluation of prediction model.

[0043] Furthermore, the final model was determined: a multi-factor logistic regression model was constructed for the five intersection candidate predictive markers obtained from the above screening:

[0044] Logit[P=HFS]=3.452+15.489*[dodecanoic acid]+8.947*[PC(14:1(9Z) / 24:1(15Z))]+2.674*[PC 34:2]+8.143*[PC 35:2]-9.803*[cis-8,11,14,17-eicosatetraenoic acid].

[0045] The [P=HFS] value was calculated, where the cutoff value of [P=HFS] was set to 0.899 (the cutoff value of [P=HFS] was calculated based on the maximum value of the Youden index). When [P=HFS]≥0.899, it indicated that the person being tested was at risk of developing an adverse reaction of hand-foot syndrome; when [P=HFS]<0.899, it indicated that the person being tested was not at risk of developing an adverse reaction of hand-foot syndrome.

[0046] The relative contents of the above five markers obtained in the Validation Set are then brought into the prediction model for capecitabine-related hand-foot syndrome adverse reactions established based on the Training Set of the present invention for analysis, and the prediction results are output.

[0047] Finally, the prediction performance of the model was evaluated in the Training Set, Validation Set, and Entire Set using indicators such as the area under the ROC curve, accuracy, precision, and recall. Figure 3 As shown, the areas under the ROC curves of the model constructed in the present invention in the Training Set, Validation Set, and Entire Set are 0.991, 0.848, and 0.954, respectively, all of which are greater than 0.8, indicating that the prediction model of the present invention can well distinguish patients with hand-foot syndrome from those without hand-foot syndrome.

[0048] As shown in Table 2, other indicators of the prediction model for the three data sets also reflect that the model has a good prediction effect, indicating that the prediction model results of the present invention are reliable and have clinical promotion value.

[0049] Table 2 Model performance of the hand-foot syndrome adverse reaction prediction model

[0050] Dataset AUROC (95% CI) TP FP TN FN Accuracy Precision Recall F1 F2 Trainingset 0.991(0.970-1.000) 28 0 15 3 0.935 1.000 0.903 0.949 0.921 Validationset 0.848(0.652-1.000) 9 2 4 2 0.765 0.818 0.818 0.818 0.818 Entire set 0.954(0.902-0.991) 37 2 19 5 0.889 0.949 0.881 0.914 0.894

Claims

1. A group of blood metabolic markers, characterized in that: Including two or more metabolic markers among dodecanoic acid, phosphatidylcholine PC (14:1(9Z) / 24:1(15Z)), phosphatidylcholine PC 34:2, phosphatidylcholine PC 35:2 and cis-8,11,14,17-eicosatetraenoic acid.

2. The set of blood metabolic markers according to claim 1, characterized in that: Including three or more metabolic markers among dodecanoic acid, phosphatidylcholine PC (14:1(9Z) / 24:1(15Z)), phosphatidylcholine PC 34:2, phosphatidylcholine PC 35:2 and cis-8,11,14,17-eicosatetraenoic acid.

3. The set of blood metabolic markers according to claim 1, characterized in that: These include dodecanoic acid, phosphatidylcholine PC (14:1(9Z) / 24:1(15Z)), phosphatidylcholine PC 34:2, phosphatidylcholine PC 35:2, and cis-8,11,14,17-eicosatetraenoic acid.

4. The set of blood metabolic markers according to claim 1, characterized in that: A multi-factor Logistic regression model is constructed based on the metabolic markers, and a prediction result is output.

5. The set of blood metabolic markers according to claim 4, characterized in that: The model is: Logit[P=HFS]=3.452+15.489*[dodecanoic acid]+8.947*[PC(14:1(9Z) / 24:1(15Z))]+2.674*[PC 34:2]+8.143*[PC 35:2]-9.803*[cis-8,11,14,17-eicosatetraenoic acid], wherein [P=HFS] represents the predictive efficacy of hand-foot syndrome based on the above biomarker combination, with a distribution range of 0-1, and the cutoff value of [P=HFS] is set to 0.899; [dodecanoic acid], [PC(14:1(9Z) / 24:1(15Z))], [PC 34:2], [PC 35:2], and [cis-8,11,14,17-eicosatetraenoic acid] respectively represent the relative contents of the metabolite markers in the blood.

6. The set of blood metabolic markers according to claim 5, characterized in that: When [P=HFS]≥0.899, it indicates that the person being tested has a risk of developing an adverse reaction of hand-foot syndrome; when [P=HFS]<0.899, it indicates that the person being tested does not have a risk of developing an adverse reaction of hand-foot syndrome.

7. Use of a group of blood metabolic markers as claimed in claim 1 in the preparation of a kit for predicting the adverse reaction of hand-foot syndrome in patients with rectal cancer after postoperative capecitabine adjuvant chemotherapy.

Images