Prediction model of breast cancer and application thereof
Through prediction models and kits based on N-glycochain omics, the problem of low accuracy of existing breast cancer diagnosis methods is solved, and high-precision breast cancer prediction and personalized treatment plans are achieved.
Patent Information
- Application Number
- CN202411789920.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-06
- Publication Date
- 2025-03-28
AI Technical Summary
The existing breast cancer diagnosis methods have low diagnostic accuracy and limited personalization, making it difficult to achieve more accurate and personalized breast cancer risk assessment and prediction.
Using a prediction model based on N-glycochain omics, the N-glycochain expression library of breast cancer patients was obtained through liquid chromatography tandem mass spectrometry technology, and N-glycochain markers with diagnostic value were screened out, and kits and models for breast cancer prediction were established.
It achieves higher breast cancer prediction accuracy (AUC>0.96), improves early diagnosis rate, and provides new ideas for personalized treatment plans.
Smart Images
Figure CN119724547A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the field of biomedicine technology, and in particular relates to a breast cancer prediction model and application thereof. Background Art
[0002] Breast cancer is one of the most common malignant tumors in women worldwide, seriously threatening women's health and quality of life. Breast cancer is a malignant tumor with highly heterogeneous biological characteristics. Although different subtypes have the same clinical symptoms and pathological manifestations, they are very different in terms of treatment effects and disease outcomes.
[0003] Currently, early diagnosis and treatment of breast cancer are crucial to improving patient survival rates. However, existing diagnostic methods have problems such as low diagnostic accuracy and limited personalization. Therefore, it is necessary to develop a new breast cancer prediction model that combines advanced statistical methods, molecular biological characteristics, and clinical data to achieve more accurate and personalized breast cancer risk assessment and prediction. The application of this prediction model will help improve the early diagnosis rate of breast cancer, the formulation of personalized treatment plans, and the improvement of patient survival rates, bringing greater clinical benefits to breast cancer patients.
[0004] At present, abnormal N-glycosylation of tissue cells has become a common feature of cell malignancy and tumor development. Changes in N-glycosylation on tumor cell surface proteins can affect tumor cell motility, intercellular signaling or metastasis, etc. The study of changes in cell surface N-glycans during the process of tissue cell carcinogenesis has very important theoretical and practical guiding significance for the elucidation of the role of sugar chains in tumor occurrence and development, the discovery of tumor sugar chain markers, and the early diagnosis and treatment of tumors and even the development of anti-tumor drugs. Summary of the invention
[0005] The purpose of the present invention is to provide a breast cancer prediction model and a breast cancer detection kit for determining the occurrence of breast cancer.
[0006] In order to achieve the above-mentioned invention object, the present invention adopts the following technical solutions:
[0007] In a first aspect, the present invention provides a breast cancer prediction model, wherein the prediction model is:
[0008] The risk score of each sample = -1.449 + 0.939 * relative content of F1H3N2S0 - 15.166 * relative content of F0H4N5S0 + 1.88 * relative content of F0H6N2S0.
[0009] Furthermore, a risk score less than zero was defined as non-breast cancer, and a risk score greater than zero was defined as breast cancer.
[0010] In a second aspect, the present invention provides a method for constructing the breast cancer prediction model, the method comprising:
[0011] Select samples from clinical patients to establish training sets and validation sets;
[0012] Extract tissue proteins, release N-glycans, and perform mass spectrometry analysis;
[0013] Analyze the relative content of N-glycans, perform ROC curve analysis, random forest analysis, and logistic regression analysis, and screen N-glycan markers;
[0014] Train and validate a predictive model for breast cancer.
[0015] Furthermore, the samples are tumor tissue and adjacent tissue.
[0016] Furthermore, according to the ROC curve analysis, 38 N-glycans with AUC greater than 0.7 were obtained: F0H3N2S0, F0H4N2S0, F0H5N2S0, F0H6N2S0, F0H7N2S0, F0H8N2S0, F0H9N2S0, F0H10N2S0, F1H5N2S0, F0H6N3S0, F1H3N2S0, F1H3N3S0, F1H4N3S0, F1H4N3S1, F1H5N4S1, F1H5N4S2, F1H6N5S0, F 1H6N5S1, F1H6N5S2, F1H7N6S0, F1H7N6S1, F1H7N6S4, F0H3N5S0, F0H4N5S0, F0H5N5S1, F1H3N5S0, F1H4N5S0, F1H5N 5S0, F1H5N5S1, F0H4N4S0, F1H4N4S0, F0H5N4S2, F0H6N5S3, F0H7N6S2, F0H7N6S3, F0H7N6S4, F1H8N7S1, F1H8N7S2.
[0017] Furthermore, 10 N-glycans were obtained through random forest analysis: F1H3N2S0, F0H4N5S0, F0H7N2S0, F1H4N5S0, F0H8N2S0, F1H4N3S1, F0H3N5S0, F0H6N2S0, F1H5N4S2, and F1H3N5S0.
[0018] In a third aspect, the present invention provides a kit for predicting breast cancer, the kit comprising:
[0019] N-glycans: relative content of F1H3N2S0, relative content of F0H4N5S0, relative content of F0H6N2S0, and the aforementioned prediction model for breast cancer.
[0020] In a fourth aspect, the present invention provides a method for predicting breast cancer using the kit, comprising the following steps:
[0021] The relative contents of three N-glycans in breast cancer patient samples were detected, and their corresponding expression values were substituted into the prediction model to predict whether it was breast cancer based on the risk score results of each sample.
[0022] Furthermore, the reagent is a mass spectrometry detection reagent for identifying the breast cancer N-glycan marker.
[0023] The positive effects of the present invention are:
[0024] The present invention provides a prediction model and application of breast cancer. The present invention uses clinical tissue as a sample, utilizes liquid chromatography tandem mass spectrometry, obtains the N-sugar chain expression library of breast cancer patients, establishes a diagnostic model based on N-omics, and provides a kit for detecting breast cancer. The prediction model includes the relative content of 3 N-sugar chains, and the high-risk score of the model predicts breast cancer. Compared with the existing traditional diagnostic methods, the present invention can more accurately predict breast cancer (AUC>0.96), which will help to improve the early diagnosis rate of breast cancer. Fewer detection indicators can be used to achieve a stable and better prediction effect, providing new ideas for personalized treatment plans for breast cancer patients.
[0025] After reading the specific embodiments of the present invention in conjunction with the accompanying drawings, other advantages and features of the present invention will become more clear. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 The present invention is a flowchart of a method for predicting a breast cancer model according to an embodiment of the present invention.
[0027] Figure 2 This is a random forest analysis diagram of N-glycans of a sample according to an embodiment of the present invention.
[0028] Figure 3 This is a diagram showing the relative contents of three N-glycans in the prediction model described in the examples of the present invention.
[0029] Figure 4 Confusion matrix of breast cancer probability for internal validation.
[0030] Figure 5 Schematic diagram of the area under the ROC curve for internal validation.
[0031] Figure 6 Breast cancer probability confusion matrix for external validation.
[0032] Figure 7 Schematic diagram of the area under the ROC curve for external validation. DETAILED DESCRIPTION
[0033] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods, but the scope of protection claimed by the present invention is not limited to the scope described in the examples.
[0034] Example 1: A method for constructing a breast cancer prediction model
[0035] Figure 1 FIG. 1 is a flow chart of a breast cancer prediction model according to an embodiment of the present invention. Figure 1 As shown, a breast cancer prediction model according to an embodiment of the present invention comprises the following steps:
[0036] (1) Sample selection
[0037] The present invention selected a total of 172 clinical tissue samples, including 86 breast cancer tissue samples and 86 normal adjacent tissue samples, established training sets and validation sets, and performed N-glycomics analysis. Specifically, 80% of the samples were used as training sets and 20% of the samples were used as validation sets. The training set included 138 clinical samples, and the validation set included 34 clinical samples.
[0038] (2) Sample preparation and mass spectrometry analysis
[0039] Take about 20 mg of tissue, add 300 μL of RIPA strong lysis solution and 3 μL PMSF, cut the tissue into pieces, vortex vigorously, lyse on ice for 30 minutes, centrifuge at 13000 rpm for 10 minutes at 4°C, and remove the supernatant. Add 4 to 6 times the volume of ice acetone to the supernatant, precipitate at -20°C overnight. Centrifuge at 5000 rpm for 10 minutes, discard the supernatant, and place the precipitate in a fume hood to dry.
[0040] Then, PNGase F enzyme digestion was performed to release N-glycans. The precipitate was dissolved in 45 μL of double distilled water, and 5 μL of 10ⅹ glycoprotein denaturation buffer was added. It was boiled at 100°C for 10 min to denature it. Then 10 μL of 10% NP-40 and 10 μL of 10ⅹ GlycoBuffer 2 reaction buffer were added, and water was added to 100 μL. Finally, 1 μL of PNGase F enzyme was added and mixed. The reaction was allowed to proceed overnight at 37°C and 130 rpm on a shaker. 3 times the volume of ice ethanol was added, and the mixture was allowed to stand at -20°C for 2 hours. The supernatant was concentrated by centrifugation to dryness, which was the N-glycan.
[0041] Then, anthranilic acid (2-AA) derivatization was performed. The N-sugar chain sample was dissolved in 20 μL of double distilled water, and then 20 μL of 2-AA labeling reagent was quickly added to the sample and reacted at 65°C in a light-proof water bath for 2 hours.
[0042] Next, mass spectrometry analysis was performed. We used an Agilent 1290 ultra-high performance liquid chromatograph in series with an LTQ OrbitrapXL mass spectrometer. On a 150mm x 2.1mm Hypercarb column, the flow rate was 0.12mL / min, and the mobile phases A and B were 10mM ammonium bicarbonate and acetonitrile (containing 20% 10mM ammonium bicarbonate), respectively. Within 40min, B increased from 12% to 28%. The scanning mode was negative ion ESI scanning, with a resolution of 60,000 and a CID collision energy of 25eV.
[0043] Finally, identification and relative quantification of N-glycans. PGC-ESI-MS / MS data were collected using Xcalibur 2.0.7 software, and then the LC-MS data were deconvoluted using Decon2LS / DeconTools software. According to ion abundance normalization, the data processed by Decon2LS / DeconTools were matched using GlycResoft software and relative quantitative analysis was performed. Finally, data matching and screening were performed using GlycoWorkBench software. Statistical analysis of the obtained N-glycan data was performed using GraphPad Prism v.5.03 software.
[0044] (3) Screening of N-glycan markers
[0045] SPSS v.22.0 was used to perform receiver operating characteristic curve analysis (ROC) and perform diagnostic analysis on statistically significant (AUC>0.7, p<0.05) N-glycan data. As shown in Table 1, 38 N-glycans with AUC greater than 0.7 were obtained according to the ROC curve analysis: F0H3N2S0, F0H4N2S0, F0H5N2S0, F0H6N2S0, F0H7N2S0, F0H8N2S0, F0H9N2S0, F0H10N2S0, F1H5N2S0, F0H6N3S0, F1H3N2S0, F1H3N3S0, F1H4N3S0, F1H4N3S1, F1H5N4S1, F1H5N4S2, F1H6N5S0, F1H6N5S1, F1H6N5S2, F1H7N6S0, F1H7N6S1, F1H7N6S4, F0H3N5S0, F0H4N5S0, F0H5N5S1, F1H3N5S0, F1H4N5S0, F1H5 N5S0, F1H5N5S1, F0H4N4S0, F1H4N4S0, F0H5N4S2, F0H6N5S3, F0H7N6S2, F0H7N6S3, F0H7N6S4, F1H8N7S1, F1H8N7S2.
[0046] Table 1 38 N-glycans with AUC greater than 0.7 in breast cancer diagnosis
[0047]
[0048]
[0049] like Figure 2 As shown, random forest analysis was performed using R language, and a total of 10 most important N-glycans for breast cancer diagnosis were obtained: F1H3N2S0, F0H4N5S0, F0H7N2S0, F1H4N5S0, F0H8N2S0, F1H4N3S1, F0H3N5S0, F0H6N2S0, F1H5N4S2, and F1H3N5S0.
[0050] (4) Establishing a prediction model
[0051] Logistic regression analysis was performed on the 10 N-glycans screened by random forest. Based on the logistic regression analysis, the coefficients of the relative contents of the three N-glycans including F1H3N2S0, F0H4N5S0, and F0H6N2S0 were determined, and a composite diagnostic scoring model for predicting T breast cancer was established. The diagnostic scoring model is:
[0052] The risk score of each sample = -1.449 + 0.939 * relative content of F1H3N2S0 - 15.166 * relative content of F0H4N5S0 + 1.88 * relative content of F0H6N2S0.
[0053] Among them, a risk score less than zero was defined as non-breast cancer, and a risk score greater than zero was defined as breast cancer.
[0054] Figure 3 The relative content differences of the three N-glycans in the normal group and the breast cancer group are shown in Table 1. The relative content of the three N-glycan markers in the breast cancer group is significantly changed compared with the normal group, and the p value is <0.0001. In one embodiment of the present invention, during the validation of the model, the training set samples are first used as internal validation. Figure 4 Figure 1. Confusion matrix for breast cancer internal validation. The internal validation model included 138 samples. The specificity of the diagnostic model was 91.3%, the sensitivity was 88.4%, the positive predictive value was 91.0%, and the negative predictive value was 88.7%. Figure 5 This is a schematic diagram of the area under the ROC curve for internal validation. The ordinate represents the sensitivity value of the diagnostic model, and the abscissa represents the specificity value of the diagnostic model. The area under the ROC curve is 0.963.
[0055] The external validation uses the validation set samples. Figure 6Figure 1. Confusion matrix of breast cancer for external validation. The external validation model included 34 samples. The specificity of the diagnostic model was 94.1%, the sensitivity was 82.4%, the positive predictive value was 93.3%, and the negative predictive value was 84.2%. Figure 7 The area under the ROC curve is 0.965 for external validation. The above results show that the diagnostic model based on N-glycans has high sensitivity and specificity and has potential application value in the clinical diagnosis of breast cancer.
[0056] The above embodiments only illustrate the technical solutions of the present invention, rather than limiting it. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for a person skilled in the art to modify the technical solutions described in the aforementioned embodiments, or to replace some of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions claimed to be protected by the present invention.
Claims
1. A breast cancer prediction model, characterized in that The prediction model is: The risk score of each sample = -1.449 + 0.939 * relative content of F1H3N2S0 - 15.166 * relative content of F0H4N5S0 + 1.88 * relative content of F0H6N2S0.
2. The breast cancer prediction model according to claim 1, characterized in that: A risk score less than zero was defined as non-breast cancer, and a risk score greater than zero was defined as breast cancer.
3. The method for constructing a breast cancer prediction model according to claim 1, characterized in that: The following steps are involved: (1) Select samples from clinical patients to establish training sets and validation sets; (2) extract tissue proteins, release N-glycans, and perform mass spectrometry analysis; (3) Analyze the relative content of N-glycans, perform ROC curve analysis, random forest analysis, and logistic regression analysis, and screen N-glycan markers; (4) Train and validate a predictive model for breast cancer.
4. The method for constructing a breast cancer prediction model according to claim 3, characterized in that: The samples are tumor tissue and adjacent tissue.
5. The method for constructing a breast cancer prediction model according to claim 3, characterized in that: According to the ROC curve analysis, 38 N-glycans with AUC greater than 0.7 were obtained: F0H3N2S0, F0H4N2S0, F0H5N2S0, F0H6N2S0, F0H7N2S0, F0H8N2S0, F0H9N2S0, F0H10N2S0, F1H5N2S0, F0H6N3S0, F1H3N2S0, F1H3N3S0, F1H4N3S0, F1H4N3S1, F1H5N4S1, F1H5N4S2, F1H6N5S0, and F1H 6N5S1, F1H6N5S2, F1H7N6S0, F1H7N6S1, F1H7N6S4, F0H3N5S0, F0H4N5S0, F0H5N5S1, F1H3N5S0, F1H4N5S0, F1H5N5 S0, F1H5N5S1, F0H4N4S0, F1H4N4S0, F0H5N4S2, F0H6N5S3, F0H7N6S2, F0H7N6S3, F0H7N6S4, F1H8N7S1, F1H8N7S2.
6. The method for constructing a breast cancer prediction model according to claim 3, characterized in that: Ten N-glycans were obtained by random forest analysis: F1H3N2S0, F0H4N5S0, F0H7N2S0, F1H4N5S0, F0H8N2S0, F1H4N3S1, F0H3N5S0, F0H6N2S0, F1H5N4S2, and F1H3N5S0.
7. A kit for predicting breast cancer, characterized in that The kit comprises: N-sugar chain: relative content of F1H3N2S0, relative content of F0H4N5S0, relative content of F0H6N2S0, and the prediction model for breast cancer according to claims 1 and 2.
8. The breast cancer prediction kit according to claim 7, characterized in that: The reagent is a mass spectrometry detection reagent for identifying the breast cancer N-sugar chain marker.