CN109988708B - System for typing a patient suffering from colorectal cancer - Google Patents
System for typing a patient suffering from colorectal cancer Download PDFInfo
- Publication number
- CN109988708B CN109988708B CN201910106934.6A CN201910106934A CN109988708B CN 109988708 B CN109988708 B CN 109988708B CN 201910106934 A CN201910106934 A CN 201910106934A CN 109988708 B CN109988708 B CN 109988708B
- Authority
- CN
- China
- Prior art keywords
- genes listed
- genes
- recurrence
- biomarker
- patients
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- C—CHEMISTRY; METALLURGY
- C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
- C12Q—MEASURING OR TESTING PROCESSES INVOLVING ENZYMES, NUCLEIC ACIDS OR MICROORGANISMS; COMPOSITIONS OR TEST PAPERS THEREFOR; PROCESSES OF PREPARING SUCH COMPOSITIONS; CONDITION-RESPONSIVE CONTROL IN MICROBIOLOGICAL OR ENZYMOLOGICAL PROCESSES
- C12Q1/00—Measuring or testing processes involving enzymes, nucleic acids or microorganisms; Compositions therefor; Processes of preparing such compositions
- C12Q1/68—Measuring or testing processes involving enzymes, nucleic acids or microorganisms; Compositions therefor; Processes of preparing such compositions involving nucleic acids
- C12Q1/6876—Nucleic acid products used in the analysis of nucleic acids, e.g. primers or probes
- C12Q1/6883—Nucleic acid products used in the analysis of nucleic acids, e.g. primers or probes for diseases caused by alterations of genetic material
- C12Q1/6886—Nucleic acid products used in the analysis of nucleic acids, e.g. primers or probes for diseases caused by alterations of genetic material for cancer
-
- C—CHEMISTRY; METALLURGY
- C12—BIOCHEMISTRY; BEER; SPIRITS; WINE; VINEGAR; MICROBIOLOGY; ENZYMOLOGY; MUTATION OR GENETIC ENGINEERING
- C12Q—MEASURING OR TESTING PROCESSES INVOLVING ENZYMES, NUCLEIC ACIDS OR MICROORGANISMS; COMPOSITIONS OR TEST PAPERS THEREFOR; PROCESSES OF PREPARING SUCH COMPOSITIONS; CONDITION-RESPONSIVE CONTROL IN MICROBIOLOGICAL OR ENZYMOLOGICAL PROCESSES
- C12Q2600/00—Oligonucleotides characterized by their use
- C12Q2600/118—Prognosis of disease development
Landscapes
- Chemical & Material Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Organic Chemistry (AREA)
- Proteomics, Peptides & Aminoacids (AREA)
- Engineering & Computer Science (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Analytical Chemistry (AREA)
- Zoology (AREA)
- Genetics & Genomics (AREA)
- Wood Science & Technology (AREA)
- Physics & Mathematics (AREA)
- Biotechnology (AREA)
- Microbiology (AREA)
- Molecular Biology (AREA)
- Hospice & Palliative Care (AREA)
- Biophysics (AREA)
- Oncology (AREA)
- Biochemistry (AREA)
- Bioinformatics & Cheminformatics (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Measuring Or Testing Involving Enzymes Or Micro-Organisms (AREA)
Abstract
The present invention provides a system for typing patients with colorectal cancer, which can show significant performance, and the survival rates without recurrence within 5 years of patients typed as high recurrence risk type and low recurrence risk type.
Description
Technical Field
The present invention relates to a system for typing a patient suffering from colorectal cancer.
Background
Colorectal cancer is the third most common cancer in the world (Ferlay et al, 2013). Approximately one third of the new cases are diagnosed as stage 2 disease, while the 5-year survival rate for stage 2 colorectal cancer patients after simple surgical resection is approximately 60-80%, and therefore, only approximately 25% of stage 2 colorectal cancer patients will require adjuvant therapy (Labianca et al, 2013). Determination of stage 2 colorectal cancer patients with high risk of recurrence after surgery can be used to select a subset of patients who may benefit from postoperative adjuvant therapy, but currently clinically widely used clinical pathological parameters of patients, such as number of lymph nodes, histological grade, deep tumor infiltration, adjacent organ invasion, do not accurately predict prognosis and accurately select patients with high risk of recurrence to receive adjuvant therapy. The present invention thus provides a method by which tumour samples from colorectal cancer patients can be typed to determine said patients with high risk of recurrence and patients with low risk of recurrence.
Disclosure of Invention
The object of the present invention is to provide a system for typing patients suffering from colorectal cancer, with which patients suffering from colorectal cancer are typed, said patients being classified into patients with a high risk of recurrence, a subgroup of patients with a high risk of recurrence being patients that should be treated post-operatively, and patients with a low risk of recurrence.
In order to realize the purpose, the technical scheme is as follows: a system for typing a patient having colorectal cancer, comprising
A data input module for inputting a first similarity value between the RNA expression level of a first biomarker in a tissue sample of a patient with colorectal cancer and the RNA expression level of the first biomarker in a tissue sample of a patient with disease relapse within five years and low cell cycle activity, and a second similarity value between the RNA expression level of the first biomarker in a tissue sample of a patient with colorectal cancer and the RNA expression level of the first biomarker in a tissue sample of a patient with no disease relapse within five years into a model calculation module, the first biomarker comprising at least 3 genes listed in table 1; inputting into the model calculation module a third similarity value between the RNA expression level of a second biomarker in a tissue sample of the patient with colorectal cancer and the RNA expression level of the second biomarker in a tissue sample of a patient with disease relapse within five years and non-low cell cycle activity, and a fourth similarity value between the RNA expression level of the second biomarker in a tissue sample of the patient with colorectal cancer and the RNA expression level of the second biomarker in a tissue sample of a patient with no disease relapse within five years, the second biomarker comprising at least 3 genes listed in table 2;
a model calculation module comprising a typing model for calculating a first and a second typing score value for a patient having colorectal cancer based on the first, second, third, fourth similarity values and the typing model; the first fractal score value = first similarity value-second similarity value, the second fractal score value = third similarity value-fourth similarity value;
a result output module, which is used for judging whether the patient has bad prognosis or good prognosis according to the first typing score value and the second typing score value of the patient with colorectal cancer, and when the first typing score value of the patient with colorectal cancer is larger than or equal to the first similarity threshold value, or/and the second typing score value of the patient with colorectal cancer is larger than or equal to the second similarity threshold value, the patient is classified as high recurrence risk and has bad prognosis; when the first typing score value of the patient with colorectal cancer < the first similarity threshold and the second typing score value of the patient with colorectal cancer < the second similarity threshold, the patient is classified as low risk of recurrence with a good prognosis.
Preferably, the first similarity threshold =0.155 and the second similarity threshold =0.076.
The present invention provides a method for typing a patient suffering from colorectal cancer, comprising the steps of:
(1) Providing an RNA sample extracted from a tumor sample of the patient, the tumor sample comprising colorectal cancer cells;
determining the RNA expression level of a biomarker in the RNA sample, wherein the biomarker comprises at least 3 genes listed in table 1;
determining a similarity value a between the RNA expression level of the biomarker and the RNA expression level of the biomarker for patients with disease relapse and low cell cycle activity within 5 years;
determining a similarity value b between the RNA expression level of the biomarker and the RNA expression level of the biomarker in patients without disease relapse within 5 years;
determining the difference e between the similarity value a and the similarity value b;
(2) Determining the RNA expression level of a biomarker in the RNA sample, wherein the biomarker comprises at least 3 genes listed in table 2;
determining a similarity value c between the RNA expression level of the biomarker and the RNA expression level of the biomarker for patients with disease relapse and non-low cell cycle activity within 5 years;
determining a similarity value d between the RNA expression level of the biomarker and the RNA expression level of the biomarker in patients without disease relapse within 5 years;
determining the difference f between the similarity value c and the similarity value d;
(3) Classifying the patient as having a poor prognosis if the difference e is above a similarity threshold or/and the difference f is above a similarity threshold; classifying the patient as having a good prognosis if the difference e is below the similarity threshold and the difference f is below the similarity threshold.
The present invention provides a system for typing a patient suffering from colorectal cancer comprising
A data input module for inputting a fifth similarity value between the RNA expression level of a third biomarker in a tissue sample of a patient with colorectal cancer and the RNA expression level of the third biomarker in a tissue sample of a patient with disease relapse within five years and low cell cycle activity, and a sixth similarity value between the RNA expression level of the third biomarker in a tissue sample of a patient with colorectal cancer and the RNA expression level of the third biomarker in a tissue sample of a patient with no disease relapse within five years into the model calculation module, the third biomarker comprising at least 3 genes listed in table 3; inputting into the model calculation module a seventh similarity value between the RNA expression level of a fourth biomarker in the tissue sample of the patient with colorectal cancer and the RNA expression level of the fourth biomarker in the tissue sample of the patient with disease relapse within five years and non-low cell cycle activity, and an eighth similarity value between the RNA expression level of the fourth biomarker in the tissue sample of the patient with colorectal cancer and the RNA expression level of the fourth biomarker in the tissue sample of the patient with disease relapse within five years, the fourth biomarker comprising at least 3 genes listed in table 4;
a model calculation module comprising a typing model for calculating a third and a fourth typing score for patients with colorectal cancer based on the fifth, sixth, seventh, eighth and typing values; the third classification score value = fifth similarity value-sixth similarity value, the fourth classification score value = seventh similarity value-eighth similarity value;
a result output module, which is used for judging whether the patient has bad prognosis or good prognosis according to the third typing score value and the fourth typing score value of the patient with colorectal cancer, and when the third typing score value of the patient with colorectal cancer is larger than or equal to a third similarity threshold value, or/and the fourth typing score value of the patient with colorectal cancer is larger than or equal to a fourth similarity threshold value, the patient is classified as high recurrence risk and has bad prognosis; when the third typing score value < the third similarity threshold for patients with colorectal cancer and the fourth typing score value < the fourth similarity threshold for patients with colorectal cancer, the patients are classified as low risk of recurrence with a good prognosis.
Preferably, the third similarity threshold =0.198 and the fourth similarity threshold = -0.003.
The present invention provides a method for typing a patient suffering from colorectal cancer, using a stable cell core gene profile, comprising the steps of:
(1) Providing an RNA sample extracted from a tumor sample of the patient, the tumor sample comprising colorectal cancer cells;
determining the RNA expression level of a biomarker in the RNA sample; wherein the biomarker comprises at least 3 genes listed in table 3;
determining a similarity value a between the RNA expression level of the biomarker and the RNA expression level of the biomarker for patients with disease relapse and low cell cycle activity within 5 years;
determining a similarity value b between the RNA expression level of the biomarker and the RNA expression level of the biomarker in patients with no disease relapse within 5 years;
determining the difference e between the similarity value a and the similarity value b;
(2) Determining the RNA expression level of a biomarker in the RNA sample; wherein the biomarker comprises at least 3 genes listed in table 4;
determining a similarity value c between the RNA expression level of the biomarker and the RNA expression level of the biomarker for patients with disease relapse and non-low cell cycle activity within 5 years;
determining a similarity value d between the RNA expression level of the biomarker and the RNA expression level of the biomarker in patients with no disease relapse within 5 years;
determining the difference f between the similarity value c and the similarity value d;
(3) Classifying the patient as having a poor prognosis if the difference e is above a similarity threshold or/and the difference f is above a similarity threshold; classifying the patient as having a good prognosis if the difference e is below the similarity threshold and the difference f is below the similarity threshold.
Preferably, the colorectal cancer comprises TNM stage II cancer according to the TNM staging system.
Preferably, the tissue sample comprises a clinically relevant sample of colorectal cancer cells or expression products of nucleic acids from colorectal cancer cells.
Preferably, the determination method of the RNA expression level comprises blotting hybridization, quantitative PCR, RNAseq sequencing and microarray analysis; the method of measuring said first, second, third, fourth, fifth, sixth, seventh or eighth similarity value comprises, but is not limited to, an euclidean distance, a manhattan distance, a chebyshev distance, a minkowski distance, a normalized euclidean distance, a mahalanobis distance, an included angle cosine, a hamming distance, a jackard distance, a correlation coefficient or an information entropy. Preferably, the method of measuring the first, second, third, fourth, fifth, sixth, seventh or eighth similarity value is a pearson correlation coefficient.
Methods for normalizing the systematic deviations of at least three genes listed in table 1 include, but are not limited to, the gene chip fRMA method, the gene chip RMA method, the RNAseq sequencing CPM method, the RNAseq sequencing FPKM method.
Normalization methods to correct for systematic variation of at least three genes listed in table 2 include, but are not limited to, the gene chip fRMA method, the gene chip RMA method, the RNAseq sequencing CPM method, the RNAseq sequencing FPKM method.
Normalization methods to correct for systematic variation of at least three genes listed in table 3 include, but are not limited to, the gene chip fRMA method, the gene chip RMA method, the RNAseq sequencing CPM method, the RNAseq sequencing FPKM method.
Normalization methods to correct for systematic variation of at least three genes listed in table 4 include, but are not limited to, the gene chip fRMA method, the gene chip RMA method, the RNAseq sequencing CPM method, the RNAseq sequencing FPKM method.
Preferably, the patient with disease recurrence and low cell cycle activity within five years is a patient who is followed up for disease recurrence and low cell cycle activity within five years, and the patient with no disease recurrence within five years is followed up for no disease recurrence within five years.
The threshold is any value that can distinguish between an RNA sample from a patient with a high risk of cancer recurrence and an RNA sample from a patient with a low risk of cancer recurrence.
Preferably, the first biomarker comprises at least 4 genes listed in table 1; more preferably, the first biomarker comprises at least 5 genes listed in table 1; more preferably, the first biomarker comprises at least 6 genes listed in table 1; more preferably, the first biomarker comprises at least 7 genes listed in table 1; more preferably, the first biomarker comprises at least 8 genes listed in table 1; more preferably, the first biomarker comprises at least 9 genes listed in table 1; more preferably, the first biomarker comprises at least 10 genes listed in table 1; more preferably, the first biomarker comprises at least 15 genes listed in table 1; more preferably, the first biomarker comprises at least 20 genes listed in table 1; more preferably, the first biomarker comprises at least 25 genes listed in table 1; more preferably, the first biomarker comprises at least 30 genes listed in table 1; more preferably, the first biomarker comprises at least 40 genes listed in table 1; more preferably, the first biomarker comprises at least 50 genes listed in table 1; more preferably, the first biomarker comprises at least 60 genes listed in table 1; more preferably, the first biomarker comprises at least 70 genes listed in table 1; more preferably, the first biomarker comprises at least 80 genes listed in table 1; more preferably, the first biomarker comprises at least 90 genes listed in table 1; more preferably, the first biomarker comprises at least 100 genes listed in table 1; more preferably, the first biomarker comprises at least 150 genes listed in table 1; more preferably, the first biomarker comprises at least 200 genes listed in table 1; most preferably, the first biomarker comprises all of the genes listed in table 1;
the second biomarker comprises at least 4 genes listed in table 2; more preferably, the second biomarker comprises at least 5 genes listed in table 2; more preferably, the second biomarker comprises at least 6 genes listed in table 2; more preferably, the second biomarker comprises at least 7 genes listed in table 2; more preferably, the second biomarker comprises at least 8 genes listed in table 2; more preferably, the second biomarker comprises at least 9 genes listed in table 2; more preferably, the second biomarker comprises at least 10 genes listed in table 2; more preferably, the second biomarker comprises at least 15 genes listed in table 2; more preferably, the second biomarker comprises at least 20 genes listed in table 2; more preferably, the second biomarker comprises at least 25 genes listed in table 2; more preferably, the second biomarker comprises at least 30 genes listed in table 2; more preferably, the second biomarker comprises at least 40 genes listed in table 2; more preferably, the second biomarker comprises at least 50 genes listed in table 2; more preferably, the second biomarker comprises at least 60 genes listed in table 2; more preferably, the second biomarker comprises at least 70 genes listed in table 2; more preferably, the second biomarker comprises at least 80 genes listed in table 2; more preferably, the second biomarker comprises at least 90 genes listed in table 2; most preferably, the second biomarker comprises all genes listed in table 2.
Preferably, the third biomarker comprises at least 4 genes listed in table 3; more preferably, the third biomarker comprises at least 5 genes listed in table 3; more preferably, the third biomarker comprises at least 6 genes listed in table 3; more preferably, the third biomarker comprises at least 7 genes listed in table 3; more preferably, the third biomarker comprises at least 8 genes listed in table 3; more preferably, the third biomarker comprises at least 9 genes listed in table 3; more preferably, the third biomarker comprises at least 10 genes listed in table 3; more preferably, the third biomarker comprises at least 15 genes listed in table 3; more preferably, the third biomarker comprises at least 20 genes listed in table 3; more preferably, the third biomarker comprises at least 25 genes listed in table 3; more preferably, the third biomarker comprises at least 30 genes listed in table 3; more preferably, the third biomarker comprises at least 40 genes listed in table 3; more preferably, the third biomarker comprises at least 50 genes listed in table 3; more preferably, the third biomarker comprises at least 60 genes listed in table 3; more preferably, the third biomarker comprises at least 70 genes listed in table 3; more preferably, the third biomarker comprises at least 80 genes listed in table 3; more preferably, the third biomarker comprises at least 90 genes listed in table 3; more preferably, the third biomarker comprises at least 100 genes listed in table 3; most preferably, the third biomarker comprises all genes listed in table 3;
the fourth biomarker comprises at least 4 genes listed in table 4; more preferably, the fourth biomarker comprises at least 5 genes listed in table 4; more preferably, the fourth biomarker comprises at least 6 genes listed in table 4; more preferably, the fourth biomarker comprises at least 7 genes listed in table 4; more preferably, the fourth biomarker comprises at least 8 genes listed in table 4; more preferably, the fourth biomarker comprises at least 9 genes listed in table 4; more preferably, the fourth biomarker comprises at least 10 genes listed in table 4; more preferably, the fourth biomarker comprises at least 15 genes listed in table 4; more preferably, the fourth biomarker comprises at least 20 genes listed in table 4; more preferably, the fourth biomarker comprises at least 25 genes listed in table 4; more preferably, the fourth biomarker comprises at least 30 genes listed in table 4; most preferably, the fourth biomarker comprises all genes listed in table 4.
Preferably, model stacking is used to increase the stability of the final parting score value. The genes in table 1 and table 2 were divided into 8 groups according to gene function: (1) cell division genes (table 1 gene function, cell cycle), (2) DNA repair genes (table 1 gene function, dnarepair), (3) epithelial mesenchymal transition genes (table 1 gene function, emt), (4) cell transfer genes (table 1 gene function, movement), (5) T cell-associated genes (table 1 gene function, tcell), (6) first 60 genes (table 1 gene function, top) with recurrence and low cell cycle activity statistics most significant, and (7) Wnt information transduction (table 2 gene function, wnt) (8) first 60 genes (table 2 gene function, top) with recurrence and non-low cell cycle activity statistics most significant. And (3) respectively calculating the typing score values of the 8 groups of genes by using a nearest centroid classification method, and finally fusing the 8 typing score values into 2 final typing score values by using a K-nearest neighbor regression method (Hechenbichler and Schliep, 2004) model. The model stacking method includes, but is not limited to, K-nearest neighbor regression, nearest centroid classification, and neural network. For each sample, if the 1 st final typing score of the sample exceeds the first similarity threshold, or the 2 nd final typing score exceeds the second similarity threshold, it is considered as a high recurrence risk sample, otherwise it is a low recurrence risk sample.
The invention also provides a method of prescribing treatment to a patient having colorectal cancer, comprising classifying the patient into a patient with high risk of recurrence and a patient with low risk of recurrence according to the system of the invention, and receiving adjuvant therapy if the patient is classified as having high risk of recurrence. Chemotherapeutic compounds include, but are not limited to, oxaliplatin, formic acid tetrahydrofolic acid, 5FU, capecitabine, eritinib, a topoisomerase 1 inhibitor. Antibody treatments include, but are not limited to, bevacizumab, cetuximab, PD-1 inhibitor oldivo, TGFBeta receptor inhibitor.
The invention provides the use of at least three genes listed in table 1, a reagent that detects the RNA expression level of at least three genes listed in table 1, at least three genes listed in table 2, or a reagent that detects the RNA expression level of at least three genes listed in table 2, in the preparation of a reagent or kit for typing a patient having colorectal cancer.
The invention provides the use of a combination of at least three genes listed in table 1 and at least three genes listed in table 2 or a combination of a reagent that detects the RNA expression level of at least three genes listed in table 1 and a reagent that detects the RNA expression level of at least three genes listed in table 2 in the preparation of a reagent or kit for typing a patient having colorectal cancer.
The invention provides the use of at least three genes listed in table 3, a reagent that detects the RNA expression level of at least three genes listed in table 3, at least three genes listed in table 4, or a reagent that detects the RNA expression level of at least three genes listed in table 4, in the preparation of a reagent or kit for typing a patient having colorectal cancer.
The invention provides the use of a combination of at least three genes listed in table 3 and at least three genes listed in table 4 or a combination of an agent that detects the RNA expression levels of at least three genes listed in table 3 and an agent that detects the RNA expression levels of at least three genes listed in table 4 in the preparation of a reagent or kit for typing a patient suffering from colorectal cancer.
Has the advantages that:
the invention provides a system for typing patients suffering from colorectal cancer, the system can show remarkable performance when being used for typing patients suffering from colorectal cancer, the typing is of a high recurrence risk type and a low recurrence risk type, the recurrence-free survival rate is remarkably different within 5 years, and the typing result is superior to the risk evaluation clinical parameters recommended by the currently used medical manuals.
Drawings
FIG. 1 is a flow chart of a method of genotyping using 4 rounds of iterative supervised machine learning and genetic function generation.
FIG. 2 is a graph of survival time for relapse in group 1 of 416 phase II colorectal cancer patients, with significantly different relapse-free survival rates in 5 years for the high and low relapse risk types, using the genes in tables 1 and 2 and the system of the invention.
FIG. 3 is a graph of survival time for relapse in group 2 of 41 stage II colorectal cancer patients, with significantly different relapse-free survival rates in 5 years for the high and low relapse risk types, using the genes in tables 1 and 2 and the system of the invention.
FIG. 4 is a graph of survival curves for time to relapse in group 1 of 416 patients with colorectal cancer in stage II using the stable cell core gene profiles in tables 3 and 4 and the system of the invention, with significantly different relapse free survival rates in 5 years for the high and low relapse risk types.
FIG. 5 is a graph of survival time for relapse in group 2 patients with stage 41 colorectal cancer, with a significant difference in relapse free survival rates between high and low relapse risk types within 5 years, using the stable cell core gene profiles in tables 3 and 4 and the system of the invention.
FIG. 6 is a plot of the recurrence risk ratio for any of the 3 to 50 genetic constructs in Table 1 for the typing method and the system of the present invention.
FIG. 7 is a plot of the recurrence risk ratio for any of the 3 to 50 genetic constructs in Table 2 for the typing method and the system of the present invention.
FIG. 8 is a plot of the recurrence risk ratio for any 3 to 50 genetic constructs in Table 3 and the system of the invention.
FIG. 9 is a plot of the recurrence risk ratio for any of the 3 to 50 genetic constructs in Table 4 for the typing method and the system of the present invention.
Detailed Description
Example 1: generation of the typing systems and methods of the invention, selection methods for patients with disease recurrence and low cell cycle activity and patients with disease recurrence and non-low cell cycle activity
Stage II colorectal cancer tumor specimens (n =416, 84 patient specimens with high risk of recurrence with recorded cancer recurrence at follow-up, 332 patient specimens with low risk of recurrence with no recorded cancer recurrence at follow-up) were used in this study. Clinical data included TNM staging of tumors, patient survival and 5-year follow-up of relapse. Gene expression data from tumor specimens, data normalized using the fRMA method. Other conventional normalization procedures may also be used.
Using gene expression data for all genes, a 4-round iterative supervised machine learning approach was used (see figure 1). For each round, a fraction of high-risk relapse specimens was compared to all low-risk relapse specimens. In these 4 rounds of repeated supervised machine learning, the high recurrence risk specimens used in each round were all different, and the selection criteria for the high recurrence risk specimens used in each round are listed in fig. 1.
The 1 st and 3 rd rounds of supervised machine learning are used to classify samples with similar characteristics and high recurrence risk, and the statistical method is to use the p-value <0.05 of the Cox genetic hazard regression of all samples in the round as the statistical standard.2 nd and 4 th rounds of supervised machine learning are used to select the final genome of the method, and 3 statistical standards are used to select genes, (1) the genes use the Cox specific risk patterns of all samples in the round as the statistical standard <0.05; (2) The difference between the median value of the gene in the current round using all high recurrence risk specimens and the median value of the gene in the current round using all low recurrence risk specimens is at least 1.2 times; (3) Genes p-value of Cox proportional risk profile of genes was recorded in 200 rounds of 10-fold cross-validation using all specimens, each cross-validation with more than 180 genes p-values. Genes selected by statistical methods were annotated for their function and subcellular localization using GO terms (The Gene Ontology Consortium, 2017).
In the 2 nd round of supervised machine learning, 65 patients with relapse phase II cancer cells were used with reduced gene expression levels of cell cycle activity relative to 332 patients without relapse phase II cancer cells with relatively low cell cycle activity, which is referred to as patients with disease relapse and low cell cycle activity within five years. As shown in Table 1, 6 groups of genes were selected in this round (1) cell division genes (cell cycle), (2) DNA repair genes (dnarepair), (3) epithelial mesenchymal transition genes (emt), (4) cell transfer genes (movement), (5) T cell-associated genes (tcell), (6) the first 60 genes (top) with recurrence and low cell cycle activity statistically most significant. To increase process stability, the cell division genome and the DNA repair genome comprise only genes encoding nuclear proteins.
In the 4 th round of supervised machine learning, the gene expression levels of the cell cycle activity of the cancer cells of the 17 patients with relapse phase II patients used were slightly increased relative to the gene expression levels of the cell cycle activity of the cancer cells of the 332 patients without relapse phase II patients, without relatively low cell cycle activity, and the patients with disease relapse and non-low cell cycle activity within five years were named. As shown in table 2, group 2 genes were selected in this round (7) Wnt signaling (Wnt) (8) with the first 60 genes (top) with the most statistically significant recurrence and non-low cell cycle activity.
To increase the stability of the final typing score value, model stacking is used. For these 8 groups of genes, each group of genes was individually assigned to calculate a typing score using the nearest centroid classification method. The 8 typing scores were fused into 2 final typing scores using a K-nearest neighbor regression algorithm (Hechenbichler and Schliep, 2004) model. And determining the optimal classification threshold of the similarity value by taking the optimal sensitivity and specificity as the standard. For each sample, if the 1 st final typing score of the sample exceeds the first similarity threshold of 0.155, or the 2 nd final typing score exceeds the second similarity threshold of 0.076, it is considered a high recurrence risk sample, otherwise it is a low recurrence risk sample.
Example 2: performance evaluation of the inventive profiling systems and methods
657 identified genes were used to construct a colorectal cancer prognostic classification, as shown in tables 1 and 2. For each stage II colorectal cancer patient, tumor specimens can be classified into 2 types: high and low risk of recurrence. In the validation of 416 patients with stage II rectal cancer in group 1 (n =416, 84 patients with high risk of recurrence had a specimen with recorded cancer recurrence at the follow-up phase, 332 patients with low risk of recurrence had no recorded cancer recurrence at the follow-up phase), both recurrence and follow-up time were positively recorded for this group of patients. Survival curve analysis showed that the survival rate for patients classified as low-risk relapse type was 97.8% [95% CI,95.7% -1] for 5 years, and 57.7% [95% CI,50.7% -65.7%) for patients classified as high-risk relapse type for 5 years. The high risk of recurrence pattern compared to the low risk of recurrence pattern was as high as HR =23.3 (p-value < 0.0001) (as in figure 2).
TABLE 1 significantly related genes
TABLE 2 significantly related genes
In group 2 validation of 41 patients with colorectal cancer II (n =41, 10 patients with high risk of recurrence with cancer recurrence recorded at follow-up, 31 patients with low risk of recurrence with no cancer recurrence recorded at follow-up). The recurrence in this group of patients was well documented, with a follow-up time of 10 years of expected value. Survival curve analysis showed that the survival rate for patients classified as low-risk relapse type was 93.8% [95% CI, 82.6% -1] for 5 years, and the survival rate for patients classified as high-risk relapse type was 50.5% [95% CI,32.0% -79.8%) for 5 years. The high risk of recurrence pattern compared to the low risk of recurrence pattern was as high as HR =11.3 (p-value = 0.00371) (see fig. 3).
The typing using the system and method of the present invention showed significant performance in both groups of patients, with significantly different relapse-free survival rates within 5 years for the high and low relapse risk types, and results superior to the risk assessment clinical parameters recommended by current medical manuals.
Example 3: performance evaluation of the typing systems and methods of the invention, using stable cell core gene profiles
The 250 stable cell core gene profiles, including 150 cell division genes, 50 DNA repair genes and the first 50 genes which were statistically most significant, as shown in tables 3 and 4, were used to construct a colorectal cancer prognostic classification. For each stage II colorectal cancer patient, tumor specimens can be classified into 2 types: high and low risk of recurrence. In the validation of the 416 patients with II-stage rectal cancer in the group 1, the relapse and the follow-up time of the patients in the group are exactly recorded. Survival curve analysis showed that the recurrence-free survival rate for patients classified as low-recurrence risk type was 93.5% [95% CI,89.7% -97.5% ], and for patients classified as high-recurrence risk type was 65.1% [95% CI, 58.7% -72.3% ], within 5 years. The high risk of recurrence pattern compared to the low risk of recurrence pattern was as high as HR =6.15 (p-value < 0.0001) (fig. 4).
TABLE 3 significantly related genes
TABLE 4 significantly related genes
In the validation of the 41 patients with stage II rectal cancer in group 2, there was an exact record of the recurrence of this group of patients, with the follow-up time being the expected value within 10 years. Survival curve analysis showed that the survival rate for patients classified as low risk for recurrence without recurrence was 87.8% [95% CI, 73.4% -1] within 5 years, and for patients classified as high risk for recurrence without recurrence was 57.2% [95% CI,38.5% -85.0%) within 5 years. The high risk of recurrence compared to the low risk of recurrence was as high as HR =4.38 (p-value = 0.0415) (fig. 5)
Using 250 stable cell core gene profiles, the typing system and method of the present invention showed significant performance in both groups of patients, with significant differences in relapse-free survival rates between high and low relapse risk types within 5 years, and results superior to risk assessment clinical parameters recommended by currently used medical manuals.
Example 4: results for a minimum of 3 genes for typing systems and methods
The number of genes used started at 2 genes, increased by 1 each time, and ended with 50 genes. For each selected number of genes (from 2 to 50), 200 rounds of simulation were performed. In each round of simulation, selected combinations of the number of genes, i.e., 200 times any combination of 2 genes, 200 times any combination of 3 genes, 200 times any combination of 4 genes, 200 times any combination of 5 genes, up to 200 times any combination of 50 genes, were selected from the genome. For each round of simulated combination of each gene number, the score for the typing method and system and the relapse risk ratio of high to low relapse risk in the patient specimens were calculated. The recurrence risk ratio for each selected basis factor was the average recurrence risk ratio for 200 rounds of simulations. The results show that only a minimum of any 3 gene combinations are required.
As can be seen from FIG. 6, any 3 to 50 of the genotyping methods constructed from the genes of Table 1 can achieve a recurrence risk ratio
As can be seen from FIG. 7, any 3 to 50 of the genotyping methods constructed from the genes in Table 2 can achieve a recurrence risk ratio
As can be seen from FIG. 8, any 3 to 50 of the genotyping methods constructed from the genes in Table 3 can achieve a recurrence risk ratio
As can be seen from FIG. 9, any 3 to 50 of the genotyping methods constructed from the genes in Table 4 can achieve a recurrence risk ratio.
Finally, it should be noted that the above embodiments are only used for illustrating the technical solutions of the present invention and not for limiting the protection scope of the present invention, and although the present invention is described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that modifications or equivalent substitutions can be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.
Claims (4)
1. Use of at least 15 genes listed in table 1, an agent that detects the RNA expression level of at least 15 genes listed in table 1, at least 15 genes listed in table 2, or an agent that detects the RNA expression level of at least 15 genes listed in table 2 in the preparation of a reagent or kit for typing a patient suffering from colorectal cancer.
2. Use of a combination of at least 15 genes listed in table 1 and at least 15 genes listed in table 2 or a combination of an agent that detects the RNA expression level of at least 15 genes listed in table 1 and an agent that detects the RNA expression level of at least 15 genes listed in table 2 in the preparation of a reagent or kit for typing a patient having colorectal cancer.
3. Use of at least 15 genes listed in table 3, an agent that detects the RNA expression level of at least 15 genes listed in table 3, at least 15 genes listed in table 4, or an agent that detects the RNA expression level of at least 15 genes listed in table 4 in the preparation of a reagent or kit for typing a patient having colorectal cancer.
4. Use of a combination of at least 15 genes listed in table 3 and at least 15 genes listed in table 4 or a combination of reagents that detect the RNA expression levels of at least 15 genes listed in table 3 and reagents that detect the RNA expression levels of at least 15 genes listed in table 4 in the preparation of a reagent or kit for typing a patient having colorectal cancer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910106934.6A CN109988708B (en) | 2019-02-01 | 2019-02-01 | System for typing a patient suffering from colorectal cancer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910106934.6A CN109988708B (en) | 2019-02-01 | 2019-02-01 | System for typing a patient suffering from colorectal cancer |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109988708A CN109988708A (en) | 2019-07-09 |
| CN109988708B true CN109988708B (en) | 2022-12-09 |
Family
ID=67129810
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910106934.6A Active CN109988708B (en) | 2019-02-01 | 2019-02-01 | System for typing a patient suffering from colorectal cancer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109988708B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113774130A (en) * | 2020-06-09 | 2021-12-10 | 碳逻辑生物科技(香港)有限公司 | Biomarkers and methods for selecting a group of chemotherapy responsive patients and uses thereof |
| CN112062827B (en) * | 2020-09-16 | 2022-10-21 | 中国人民解放军军事科学院军事医学研究院 | Application of CEP55 protein in regulation of cilia de-assembly and preparation of cilia-related disease model |
| CN113436741B (en) * | 2021-07-16 | 2023-02-28 | 四川大学华西医院 | Lung cancer recurrence prediction method based on tissue specific enhancer region DNA methylation |
| CN116870139A (en) * | 2023-09-06 | 2023-10-13 | 暨南大学附属第一医院（广州华侨医院） | Application of LASP1 protein in preparation of medicine for treating spinal cord injury repair |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8105777B1 (en) * | 2008-02-13 | 2012-01-31 | Nederlands Kanker Instituut | Methods for diagnosis and/or prognosis of colon cancer |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2009002175A1 (en) * | 2007-06-28 | 2008-12-31 | Agendia B.V. | A method of typing a sample comprising colorectal cancer cells |
| EP2195658A2 (en) * | 2007-09-28 | 2010-06-16 | Royal College of Surgeons in Ireland | A method of assessing colorectal cancer status in an individual |
| AU2015202173A1 (en) * | 2007-10-05 | 2015-05-14 | Pacific Edge Limited | Proliferation signature and prognosis for gastrointestinal cancer |
| GB0720113D0 (en) * | 2007-10-15 | 2007-11-28 | Cambridge Cancer Diagnostics L | Diagnostic, prognostic and predictive testing for cancer |
| EP2202320A1 (en) * | 2008-12-24 | 2010-06-30 | Agendia B.V. | Methods and means for typing a sample comprising colorectal cancer cells |
| US20140031251A1 (en) * | 2010-11-03 | 2014-01-30 | H. Lee Moffitt Cancer Center And Research Institute, Inc. | Methods of classifying human subjects with regard to cancer prognosis |
| EP3009842B1 (en) * | 2014-09-26 | 2019-09-04 | Sysmex Corporation | Method for supporting diagnosis of risk of colorectal cancer recurrence, program and computer system |
| CN109938695A (en) * | 2019-03-08 | 2019-06-28 | 度特斯(大连)实业有限公司 | A kind of human body diseases Risk Forecast Method and equipment based on heterogeneous degree index |
-
2019
- 2019-02-01 CN CN201910106934.6A patent/CN109988708B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8105777B1 (en) * | 2008-02-13 | 2012-01-31 | Nederlands Kanker Instituut | Methods for diagnosis and/or prognosis of colon cancer |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109988708A (en) | 2019-07-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109988708B (en) | System for typing a patient suffering from colorectal cancer | |
| JP2021061844A (en) | Using size and number aberrations in plasma dna for detecting cancer | |
| JP2022521492A (en) | An integrated machine learning framework for estimating homologous recombination defects | |
| US20090062144A1 (en) | Gene signature for prognosis and diagnosis of lung cancer | |
| WO2019023517A2 (en) | Genomic sequencing classifier | |
| CN110305964B (en) | Tool for prognostic recurrence risk prediction marker of prostate cancer patient and establishment of risk assessment model of tool | |
| KR20170125044A (en) | Mutation detection for cancer screening and fetal analysis | |
| US8030060B2 (en) | Gene signature for diagnosis and prognosis of breast cancer and ovarian cancer | |
| CN110273003B (en) | Marker tool for prognosis recurrence detection of papillary renal cell carcinoma patient and establishment of risk assessment model thereof | |
| US20160102359A1 (en) | Genetic marker for early breast cancer prognosis prediction and diagnosis, and use thereof | |
| Chen et al. | DNA methylation-based classification and identification of renal cell carcinoma prognosis-subgroups | |
| WO2013086352A1 (en) | Prostate cancer associated circulating nucleic acid biomarkers | |
| EP2419540B1 (en) | Methods and gene expression signature for assessing ras pathway activity | |
| AU2005312081A1 (en) | Methods and systems for prognosis and treatment of solid tumors | |
| Zhao et al. | Identification of pan-cancer prognostic biomarkers through integration of multi-omics data | |
| EP3950960A1 (en) | Dna methylation marker for predicting recurrence of liver cancer, and use thereof | |
| CN109929934B (en) | Application of immune related gene in kit and system for colorectal cancer prognosis | |
| Zhu et al. | DNA methylation biomarkers for the occurrence of lung adenocarcinoma from TCGA data mining | |
| CN115820860A (en) | Method for screening non-small cell lung cancer marker based on methylation difference of enhancer, marker and application thereof | |
| Belvedere et al. | A computational index derived from whole-genome copy number analysis is a novel tool for prognosis in early stage lung squamous cell carcinoma | |
| CN114829624A (en) | Genomic scar assay and related methods | |
| CN115472217A (en) | System for predicting pancreatic cancer patient prognosis | |
| JP2023529759A (en) | How to predict cancer progression | |
| Ma et al. | Optimizing the Prognostic Model of Cervical Cancer Based on Artificial Intelligence Algorithm and Data Mining Technology | |
| US10240206B2 (en) | Biomarkers and methods for predicting benefit of adjuvant chemotherapy |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information | ||
| CB03 | Change of inventor or designer information |
Inventor after: Sun Tian Inventor after: Wu Rujia Inventor before: Sun Tian |
|
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP03 | Change of name, title or address | ||
| CP03 | Change of name, title or address |
Address after: A133, Building B, No. 1 Huanzhen East Road South, Beijiao Community, Beijiao Town, Shunde District, Foshan City, Guangdong Province, 528000 (Residence Application) Patentee after: Carbon Logic Biotechnology (Foshan) Co.,Ltd. Address before: 528,400 Card 502, Floor 5, Shumao Building, No. 6, Xiangxing Road, Torch Development Zone, Zhongshan City, Guangdong Province Patentee before: Carbon Logic Biotechnology (Zhongshan) Co.,Ltd. |