阅读说明：本技术 一种化疗相关基因表达特征在预测胰腺癌预后中的应用 (Application of chemotherapy-related gene expression characteristics in prediction of pancreatic cancer prognosis ) 是由 浦宁 吴文川 陈强达 张继成 殷瀚霖 楼文晖 于 2021-06-29 设计创作，主要内容包括：本发明涉及一种化疗相关基因表达特征在预测胰腺癌预后中的应用,属于生物医药技术领域。本发明提供了一种化疗相关基因表达特征在预测胰腺癌化疗反应应答、免疫细胞浸润及预测模型中的应用,通过检测胰腺癌患者术后肿瘤组织中BCHE,ADH1A和TSN4的基因表达水平,计算相关基因表达特征评分作为客观指标反映胰腺癌化疗药物敏感性、肿瘤纯度、肿瘤免疫细胞浸润水平及潜在的肿瘤生物学行为,用于指导胰腺癌患者术后化疗及免疫治疗,推动胰腺癌个体化治疗；同时,本发明基于化疗相关基因表达特征行胰腺癌预后风险分层,具有更好的预测效能,使胰腺癌患者获得更多的临床净收益,应用于临床准确判断胰腺癌患者术后生存时间。(The invention relates to application of chemotherapy-related gene expression characteristics in prediction of pancreatic cancer prognosis, and belongs to the technical field of biological medicines. The invention provides application of chemotherapy-related gene expression characteristics in prediction of pancreatic cancer chemotherapy response, immune cell infiltration and prediction models, wherein by detecting gene expression levels of BCHE, ADH1A and TSN4 in postoperative tumor tissues of pancreatic cancer patients, related gene expression characteristic scores are calculated and used as objective indexes to reflect drug sensitivity, tumor purity, tumor immune cell infiltration level and potential tumor biological behaviors of pancreatic cancer chemotherapy, so that the chemotherapy and immunotherapy of the pancreatic cancer patients after operation are guided, and individualized treatment of pancreatic cancer is promoted; meanwhile, the method carries out pancreatic cancer prognosis risk stratification based on the expression characteristics of chemotherapy-related genes, has better prediction efficiency, enables pancreatic cancer patients to obtain more clinical net benefits, and is applied to clinically and accurately judging the postoperative survival time of the pancreatic cancer patients.)

1. The application of chemotherapy-related gene expression characteristics in predicting pancreatic cancer chemotherapy response and immune cell infiltration is characterized in that: the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression.

2. The use of a chemotherapy-associated gene expression signature in the prediction of pancreatic cancer chemotherapy response and immune cell infiltration according to claim 1, wherein: the gene tissue expression condition comprises a gene expression value, the calculation method of the gene expression value is a real-time fluorescence quantitative PCR result value of a target gene, and log2 transformation is carried out after the result value of an internal reference gene GAPDH is standardized.

3. The use of a chemotherapy-associated gene expression signature in the prediction of pancreatic cancer chemotherapy response and immune cell infiltration according to claim 1, wherein: the gene tissue expression condition comprises chemotherapy-related gene expression feature score calculation, wherein the feature score calculation formula is | [ (-0.60912) × BCHE expression value ] + [ (-1.61995) × ADH1A expression value ] + [0.01775 × TNS4 expression value ] |; and dividing into a high-scoring group and a low-scoring group according to the feature scores.

4. The application of a prediction model based on chemotherapy-related gene expression characteristics in predicting postoperative chemotherapy response of pancreatic cancer patients is characterized in that: the prediction model comprises a factor input module and a prediction result output module; the factor input module comprises the feature score of claim 3; the prediction result output module comprises the probability of relapse-free survival of the pancreatic cancer patient after the operation of chemotherapy.

5. Use of a prediction model based on expression signatures of chemotherapy-associated genes for predicting the response to postoperative chemotherapy in pancreatic cancer patients according to claim 4, wherein: the construction method of the prediction model comprises the steps of obtaining independent prediction factors of relapse-free survival time through multi-factor proportional risk regression analysis as input factors of the model, calculating variable partial regression coefficients through Cox regression, and imaging the prediction model into a Nomogram model.

6. The application of chemotherapy-related gene expression characteristics in preparing a detection kit for predicting pancreatic cancer chemotherapy response is characterized in that: the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression.

7. An application of chemotherapy-related gene expression characteristics in preparation of a detection kit for predicting pancreatic cancer immune cell infiltration is characterized in that: the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression.

8. The application of chemotherapy-related gene expression characteristics in preparing a detection kit for predicting the recurrence-free survival probability of postoperative adjuvant chemotherapy of pancreatic cancer is characterized in that: the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression.

Technical Field

The invention relates to application of chemotherapy-related gene expression characteristics in prediction of pancreatic cancer prognosis, and belongs to the technical field of biological medicines.

Background

Recent global epidemiological surveys have shown that ductal adenocarcinoma of the pancreas is one of the most lethal and aggressive malignancies with a 5-year survival rate of less than 9%. Most patients lose the opportunity to receive radical resection when diagnosed due to the occult and asymptomatic onset. Only < 20% of patients have an opportunity to receive surgical resection, but disappointing is that their 5-year survival rate is also only 20-30%. In pancreatic cancer treatment, except for surgical resection, postoperative adjuvant chemotherapy also shows certain therapeutic effects, and the modified FOLFIRINOX or gemcitabine-based regimen is widely accepted as the first-line method of postoperative treatment. However, there are no sufficiently sensitive and specific indexes and models clinically for accurately predicting the postoperative chemotherapy response and the immune cell infiltration condition of pancreatic cancer patients so as to guide the formulation of treatment strategies and improve the survival rate of patients after radical operation.

Gemcitabine is a nucleoside drug that competes for DNA synthesis and inhibits tumor proliferation. An early clinical study comparing gemcitabine with 5-fluorouracil in the treatment of advanced pancreatic cancer shows that gemcitabine has better clinical benefit in advanced PDACs. Subsequent randomized controlled trials showed that gemcitabine has a high value for postoperative adjuvant treatment of pancreatic cancer, with median disease-free survival increasing from 6.7 months to 13.4 months and overall survival increasing from 20.2 months to 22.1 months. Studies have shown that improved FOLFINIROX has a better effect after pancreatic cancer surgery, with OS up to 54.5 months. Gemcitabine and FOLFIRINOX-based chemotherapy regimens greatly prolonged the survival of pancreatic cancer patients, however, not all patients benefit from postoperative adjuvant chemotherapy. Given that short-term or long-term toxic complications after chemotherapy are inevitable, screening for predictive biomarkers will help select beneficial populations, which will be of greater clinical value. The current research finds that pancreatic cancer tissues have obvious connective tissue proliferative response and high genetic heterogeneity. With the development of gene chips and high-throughput sequencing technologies, gene expression profiles may provide more useful information for chemotherapy sensitivity and survival of pancreatic cancer patients. In pancreatic cancer, malignant tumor cells are surrounded by an abundant stroma, consisting mainly of cancer-associated fibroblasts, extracellular matrix, and immune cells. The biological behavior of pancreatic cancer cells may be exerted by the matrix by promoting their proliferation and survival in harsh environments. To date, several genes, tumor infiltrating cells, or interstitial/immune scores have been shown to be of clinical interest in predicting survival or chemotherapy response.

In previous reports, the effect of a single gene or non-coding RNA in pancreatic cancer chemotherapy resistance and the prediction of the chemotherapy curative effect thereof are widely reported. However, due to the high heterogeneity of pancreatic cancer, it is not complete to predict post-operative chemosensitivity with only one biomarker with sufficient specificity. Therefore, there is a need in the art to obtain a gene signature associated with chemotherapy response to predict the response and survival time of adjuvant chemotherapy after pancreatic cancer surgery.

Disclosure of Invention

The invention aims to overcome the defects in the prior art, provides application of chemotherapy-related gene expression characteristics in predicting pancreatic cancer chemotherapy response and immune cell infiltration, determines molecular characteristics which can be detected in tumor tissues, reflect the sensitivity of chemotherapy response and immune subgroup infiltration after pancreatic cancer surgery and judge adverse prognosis risks, can be used for screening patients effective in chemotherapy or immune treatment, and improves treatment effect, thereby promoting individualized accurate treatment. The prediction model based on the chemotherapy-related gene expression characteristics is provided, can quantitatively and accurately predict the non-recurrence survival probability of the pancreatic cancer patient after postoperative adjuvant chemotherapy, and can be used for identifying the high-risk patient to perform key monitoring and timely replacing a treatment scheme, thereby improving the prognosis of the patient. The invention aims to solve the technical problems of how to screen patients effective in pancreatic cancer chemotherapy or immunotherapy, improve the treatment effect, promote individualized accurate treatment and how to identify high-risk patients.

In order to solve the problems, the technical scheme adopted by the invention is to provide application of chemotherapy-related gene expression characteristics in predicting pancreatic cancer chemotherapy response and immune cell infiltration, wherein the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression conditions.

Preferably, the expression condition of the gene tissue comprises a gene expression value, and the calculation method of the gene expression value is a real-time fluorescent quantitative PCR result value of a target gene, and is converted into log2 after being normalized by a result value of a reference gene GAPDH.

Preferably, the gene tissue expression condition comprises chemotherapy-related gene expression characteristic score calculation, and the characteristic score calculation formula is | [ (-0.60912) × BCHE expression value ] + [ (-1.61995) × ADH1A expression value ] + [0.01775 × TNS4 expression value ] |; and dividing into a high-scoring group and a low-scoring group according to the feature scores.

The invention provides application of a prediction model based on chemotherapy-related gene expression characteristics in predicting postoperative chemotherapy response of pancreatic cancer patients, wherein the prediction model comprises a factor input module and a prediction result output module; the factor input module comprises the feature score; the prediction result output module comprises the probability of relapse-free survival of the pancreatic cancer patient after the postoperative chemotherapy; the calculation method of the postoperative chemotherapy recurrence-free survival probability of the pancreatic cancer patient comprises the step of calculating the probability scale axis value corresponding to the position of the feature score in the corresponding prediction model.

Preferably, the construction method of the prediction model comprises the steps of obtaining independent prediction factors of recurrence-free survival time through multi-factor proportional risk regression analysis as input factors of the model, calculating variable partial regression coefficients through Cox regression, and imaging the prediction model into a Nomogram model.

The invention provides application of chemotherapy-related gene expression characteristics in preparation of a detection kit for predicting pancreatic cancer chemotherapy response, wherein the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression conditions.

The invention provides application of chemotherapy-related gene expression characteristics in preparation of a detection kit for predicting pancreatic cancer immune cell infiltration, wherein the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression conditions.

The invention provides application of chemotherapy-related gene expression characteristics in preparation of a detection kit for predicting postoperative adjuvant chemotherapy relapse-free survival probability of pancreatic cancer, wherein the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression conditions.

Compared with the prior art, the invention has the following beneficial effects:

the invention can simply and rapidly reflect the individual tumor immune cell infiltration degree and the response level of postoperative adjuvant chemotherapy reaction by detecting the feature score level of chemotherapy-related gene expression in the pancreatic cancer postoperative tumor tissue; the chemo-related gene expression characteristic-based Nomogram recurrence prediction model provided by the invention can quantitatively and accurately evaluate the probability of the postoperative adjuvant chemotherapy recurrence-free survival time of pancreatic cancer patients, and preliminarily and effectively evaluate the risk of poor prognosis of the patients, so that valuable references can be timely provided for postoperative treatment strategies in clinical practice, and the model has a high guiding significance on selection of chemical and immunotherapy and is beneficial to individualized treatment of pancreatic cancer.

Drawings

Fig. 1 is an analysis diagram of pancreatic cancer RNA sequencing data and clinical chemotherapy data of a TCGA public database on chemotherapy-resistant and sensitive group differential genes in an embodiment of the present invention (a diagram is a volcano diagram of all genes of a chemotherapy-resistant drug group and a sensitive group after a pancreatic cancer operation, each scatter point on the diagram is each gene, a down-regulated gene with a significant difference is in an upper left quadrant, an up-regulated gene with a significant difference is in an upper right quadrant, and a B diagram is an expression pattern hierarchical clustering analysis diagram of the first 50 significantly-differentiated genes, where a darker red color represents a higher expression, and a darker blue color represents a lower expression).

FIG. 2 is an analysis chart and a verification chart for establishing expression characteristic scores of chemotherapy-associated genes in the examples of the present invention (A and B are Lasso-pendizedCox analyses of differential genes selected from 25 survival-associated differential genes, i.e., TNS4, WFDC3, IHH, AGTR1, CHMP4BP1, BCHE, ITPKA, MIR3131, RN7SL138P, CRCT1, FER1L6, CAPN9, ADH1A, CLDN4, SULT1B1, LINC02593, CCK, SPDEF, ITM2A, ADHMA 7A, DNAN 4, TRIM7, ADRB 8, IGLV3-10 and SDR16C5, each of A represents a gene, the position indicated by the middle dotted line in B represents the least gene, the graph of the TCGA expression score calculated from the comparison of the expression characteristic scores of the chemotherapeutic drug resistance group and TCE sensitivity group, and TCGA sensitivity score calculated from the TCE group-sensitive gene group-group analysis, panel F is a comparative analysis of TNS4 gene expression between chemotherapy resistant and sensitive groups in the TCGA cohort).

Fig. 3 is an analysis graph of chemotherapy-related gene expression signature scores in predicting postoperative adjuvant chemotherapy response and survival prognosis established in the example of the present invention (a graph is a total survival time curve of patients with high expression signature scores and low expression signature scores in TCGA cohort, B graph is a ROC curve analysis of chemotherapy-related gene expression signature scores in TCGA cohort for response to postoperative adjuvant chemotherapy, C graph is a ROC curve analysis of chemotherapy-related gene expression signature scores in TCGA cohort for total survival level of postoperative adjuvant chemotherapy, and D graph is a incorporated ROC time-to-live curve of patients with high chemotherapy-related gene expression signature scores and low chemotherapy-related gene expression signature scores in central pancreatic cancer patients).

FIG. 4 is a diagram of a preliminary analysis of the extent of immune cell infiltration by CIBERSORT algorithm based on RNA sequencing data of pancreatic cancer in TCGA public databases in an example of the present invention (A is a graph showing a pattern of the proportion of different immune cell infiltrations in the taken-in case tissues, and B is a graph showing a comparative analysis of the proportion of immune cell infiltrations in the chemotherapy-resistant drug group and the sensitive group in the TCGA cohort).

FIG. 5 is an example of immunohistochemical staining of the degree of infiltration of immune cell subsets in pancreatic cancer tissues in accordance with an embodiment of the present invention (case 1 shows high infiltration of CD4+ T cells, high infiltration of CD8+ T cells, medium infiltration of CD11C + dendritic cells, high infiltration of CD68+ macrophages, and low infiltration of MPO + neutrophils, and case 2 shows low infiltration of CD4+ T cells, low infiltration of CD8+ T cells, medium infiltration of CD11C + dendritic cells, medium infiltration of CD68+ macrophages, and high infiltration of MPO + neutrophils).

FIG. 6 is a diagram of a statistical analysis of the degree of infiltration of immune cell subsets in pancreatic cancer tissues according to an embodiment of the present invention (A is a comparison of the high chemotherapy-related gene expression signature score and the infiltration level of CD4+ T cells in the low chemotherapy-related gene expression signature score tissue of the central pancreatic cancer patient, B is a comparison of the high chemotherapy-related gene expression signature score and the infiltration level of CD8+ T cells in the low chemotherapy-related gene expression signature score tissue of the central pancreatic cancer patient, a comparison of the high chemotherapy-related gene expression signature score and the infiltration level of CD11C + dendritic cells in the low chemotherapy-related gene expression signature score tissue of the central pancreatic cancer patient, and a comparison of the high chemotherapy-related gene expression signature and the CD68+ macrophage infiltration level in the low chemotherapy-related gene expression signature score tissue of the central pancreatic cancer patient, high chemotherapy-associated gene expression signature scores in the present central pancreatic cancer patients for inclusion compared to the level of infiltration of MPO + neutrophils in the low chemotherapy-associated gene expression signature score tissues).

FIG. 7 is a Nomogram prognostic prediction model of RFS at 6, 12 and 18 months after surgery in pancreatic cancer patients receiving adjuvant chemotherapy constructed in an embodiment of the present invention.

Detailed Description

In order to make the invention more comprehensible, preferred embodiments are described in detail below with reference to the accompanying drawings:

the invention provides application of chemotherapy-related gene expression characteristics in predicting pancreatic cancer chemotherapy response and immune cell infiltration, wherein the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression conditions. The gene tissue expression condition comprises a gene expression value, the calculation method of the gene expression value is a real-time fluorescence quantitative PCR result value of a target gene, and log2 transformation is carried out after the result value of an internal reference gene GAPDH is standardized. The gene tissue expression condition comprises the calculation of chemotherapy-related gene expression feature score, wherein the feature score is calculated according to the formula | [ (-0.60912) × BCHE expression value ] + [ (-1.61995) × ADH1A expression value ] + [0.01775 × TNS4 expression value ] |; and dividing into a high-scoring group and a low-scoring group according to the feature scores.

The invention provides application of a prediction model based on chemotherapy-related gene expression characteristics in predicting postoperative chemotherapy response of pancreatic cancer patients, wherein the prediction model comprises a factor input module and a prediction result output module; the factor input module comprises the feature scores; the prediction result output module comprises the probability of relapse-free survival of the pancreatic cancer patient after the operation of chemotherapy. The construction method of the prediction model comprises the steps of obtaining independent prediction factors of relapse-free survival time through multi-factor proportional risk regression analysis as input factors of the model, calculating variable partial regression coefficients through Cox regression, and imaging the prediction model into a Nomogram model.

The invention provides application of chemotherapy-related gene expression characteristics in preparation of a detection kit for predicting pancreatic cancer chemotherapy response, wherein the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression conditions.

The invention provides application of chemotherapy-related gene expression characteristics in preparation of a detection kit for predicting pancreatic cancer immune cell infiltration, wherein the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression conditions.

The invention provides application of chemotherapy-related gene expression characteristics in preparation of a detection kit for predicting postoperative adjuvant chemotherapy relapse-free survival probability of pancreatic cancer, wherein the chemotherapy-related gene expression characteristics comprise BCHE, ADH1A and TNS4 gene tissue expression conditions.

Aiming at the defects in the prior art, the primary purpose of the invention is to provide the application of chemotherapy-related gene expression characteristics in predicting pancreatic cancer chemotherapy response and immune cell infiltration, and to determine a molecular characteristic which can be detected in tumor tissues and can reflect the sensitivity of chemotherapy response and immune subgroup infiltration after pancreatic cancer surgery and judge adverse risk of prognosis, so that the molecular characteristic can be used for screening patients effective in chemotherapy or immune treatment, improving the treatment effect and promoting individualized precise treatment.

The second purpose of the invention is to provide a prediction model based on chemotherapy-related gene expression characteristics, namely a Nomogram prediction model capable of quantitatively and accurately predicting the recurrence-free survival probability of pancreatic cancer patients after postoperative adjuvant chemotherapy, which can be used for identifying high-risk patients to perform key monitoring and timely replacing treatment schemes, thereby improving the prognosis of patients.

In order to achieve the above primary object, the solution of the present invention is: provides an application of chemotherapy-related gene expression characteristics in predicting pancreatic cancer chemotherapy response and immune cell infiltration. The establishment of the expression characteristics of the chemotherapy-related genes comprises the following steps:

(1) screening pancreatic cancer sequencing data in the TCGA database, bringing the pancreatic cancer sequencing data into postoperative auxiliary chemotherapy cases meeting the conditions, and dividing the pancreatic cancer sequencing data into a chemotherapy sensitive group and a chemotherapy resistant group.

(2) And determining differential genes (DEGs) between chemotherapy sensitive groups and drug resistant groups by using an R language Limma package under the condition that | Log2fold change | is >1and p is less than 0.05.

(3) Using the CIBERSORT algorithm

(https:// ciberstart. stanford. edu/index. php), tumor infiltrating immune cell composition was analyzed.

(4) Performing single-factor Cox survival analysis by using the obtained DEGs, setting p to be less than 0.05, and determining survival related DEGs; and finally determining DEGs related to the independence of chemotherapy reaction and survival prognosis by utilizing Lasso-pendized Cox analysis and multifactor Cox survival analysis.

In the step (4), it is found through research that the finally obtained DEGs are butyrylcholinesterase (bche), alcohol dehydrogenase1A (ADH1A) and tenin 4(TNS4), and the final name of the gene feature is composed of the initials of three gene names, namely, the characteristic named as "BAT" - "BAT". According to the regression coefficient of the multi-factor Cox survival analysis, the 'bat' feature score calculation formula is determined as follows: | [ (-0.60912) × BCHE expression value ] + [ (-1.61995) × ADH1A expression value ] + [0.01775 × TNS4 expression value ] |.

The bat feature verification comprises the following steps:

(1) and collecting surgical excision specimens of pancreatic cancer patients, wherein one part of the specimens is used for extracting RNA, the other part of the specimens is fixed by paraformaldehyde, and a tissue microarray chip is constructed by paraffin embedding.

(2) Analyzing the expression of BCHE, ADH1A and TNS4 by real-time fluorescence quantitative PCR of sample RNA; after paraffin removal from the tissue microarray sections, the sections were incubated overnight with antibodies against CD4, CD8, CD11C, MPO, CD68, and then with a secondary antibody conjugated with horseradish peroxidase, followed by immunohistochemical staining, i.e., nuclear staining with Hariss hematoxylin.

(3) Dividing the final infiltration level of each immune cell subgroup in the tissue of the pancreatic cancer patient into high infiltration and low infiltration according to the immunohistochemical staining result, and analyzing and statistically comparing the prognosis of the patient with high infiltration and low infiltration according to the clinical follow-up information of the patient.

(4) The expression of BCHE, ADH1A and TNS4 measured by real-time fluorescent quantitative PCR is standardized by the result value of an internal reference gene GAPDH, then log2 is converted, the characteristic score of bat is calculated and further divided into high and low groups, and the relapse-free survival time of the high and low groups of patients is analyzed and statistically compared according to the clinical follow-up information of the patients to reflect the response condition of postoperative adjuvant chemotherapy reaction; meanwhile, the correlation between the characteristic score of the bat and the cell infiltration of the tumor immune subpopulation is evaluated by using independent sample t test or rank sum test.

In the step (3), the method for evaluating the infiltration level of the tissue immune cells is a positive cell counting method;

by calculating the interstitial area (mm) on the tissue section²) And the number of different positive immune cells in the region, using the following formula to calculate: total immune cell count/interstitial area (mm)²)；

The optimal cut-off value for each immune subpopulation was calculated according to the john's index of the ROC curve, with high infiltrates defined above the cut-off value and low infiltrates defined below the cut-off value.

In the step (4), through research, the sizes of the characteristic scores of the bat for pancreatic cancer are determined and calculated, the patients are divided into high-score groups and low-score groups, the early-stage recurrence risk of the patients after postoperative adjuvant chemotherapy can be effectively judged, and meanwhile, the tumors with high characteristic scores of the bat for pancreatic cancer are proved to have lower CD4+ and CD8+ T cell infiltration.

In order to achieve the second objective, the solution of the invention is:

an application of a prediction model based on bat characteristics in predicting the relapse-free survival time of postoperative adjuvant chemotherapy of pancreatic cancer; in effect, the "bat" feature is the "bat" feature described above. The prediction model is composed of a factor input module and a prediction result output module. Wherein, the factor input module includes: "bat" feature scoring; the prediction result output module comprises: probability of survival without recurrence in pancreatic cancer patients after surgery.

Wherein, the calculation methods of the probability of survival without relapse after 6 months, 12 months and 18 months of operation are respectively as follows: and (3) a probability scale axis value corresponding to the position of the bat feature score in the corresponding prediction model.

The construction method of the prediction model comprises the following steps:

and respectively obtaining recurrence-free survival independent prediction factors as input factors of the model through multi-factor Cox regression analysis, calculating a partial regression coefficient of each variable through Cox regression, and imaging the prediction model into a Nomogram model.

Example (b):

1. study subjects:

the subjects of this example were 29 patients with pancreatic cancer, and the inclusion and exclusion criteria were:

(1) no history of other malignant tumors;

(2) the cancer treatment is not received before the operation, and the metastasis is not found before the operation;

(3) no signs of infection or other inflammation prior to surgery;

(4) radical excision;

(5) the post-operative pathology was clearly diagnosed as pancreatic ductal adenocarcinoma.

2. The research method comprises the following steps:

(1) collecting a tumor tissue specimen which is excised in an operation of a pancreatic cancer patient, and extracting mRNA of a part of tissues by using carbon tetrachloride, isopropanol and absolute ethanol solution; a portion of the tissue was fixed with paraformaldehyde and embedded in paraffin to construct a tissue microarray chip.

(2) Collecting clinical pathology information and postoperative follow-up data of pancreatic cancer patients

Conventional clinical pathological information includes age, sex, carcinoembryonic antigen (CEA), carbohydrate antigen 19-9(CA19-9), carbohydrate antigen 125(CA125), carbohydrate antigen 50(CA50) levels. Postoperative pathological reports include tumor size, location, differentiation, nerve invasion, vascular invasion and lymphatic metastasis. The postoperative adjuvant treatment conditions include chemotherapy regimen, whether radiotherapy is needed, etc.

Follow-up was performed 1 time every 3 months in 1 year, followed by 1 time every 3-6 months in 2 years, and then 1 time per year. The follow-up contents comprise blood routine, liver function examination, serum tumor marker detection, abdominal ultrasound and chest radiography. Enhanced CT or magnetic resonance imaging is performed on patients suspected of relapse. Recurrence-free survival (RFS) is defined as the time interval from the date of surgery to the date of recurrence (or last follow-up).

(3) Establishment of bat characteristics related to postoperative adjuvant chemotherapy response and survival prognosis of pancreatic cancer

By obtaining 182 pancreatic cancer patient tissue sequencing data from a TCGA public database, after matching with clinical postoperative chemotherapy and follow-up data, 26 pancreatic cancer tissues are finally confirmed to be included in a chemotherapy resistant medicine group, and 24 pancreatic cancer tissues are finally confirmed to be included in a chemotherapy sensitive group. Using the Limma package in the R language, setting conditions of | Log2fold change | >1and p <0.05, and determining Differential Expressed Genes (DEGs) between groups. Performing unifactor Cox survival analysis by using the obtained DEGs, setting p to be less than 0.05, and determining survival related DEGs; and finally determining DEGs related to the independence of chemotherapy reaction and survival prognosis by utilizing Lasso-pendized Cox analysis and multi-factor Cox survival analysis, and determining a 'bat' feature score calculation formula according to regression coefficients of the multi-factor Cox survival analysis.

(4) Real-time fluorescent quantitative PCR analysis of relativity between bat characteristics and chemotherapy response reaction

Converting the extracted mRNA into cDNA by using a reverse transcription kit, and designing primer sequences BCHE, ADH1A and TNS4 gene primers as follows: F-BCHE: TCCATAGTGAAACGGTGGGC, R-BCHE: AGGCCAGCTTGTGCTATTGT, respectively; F-ADH 1A: AAGTATCCGTACCATTCTGATGTTT, R-ADH 1A: TCTTTGGAAAGCCCCCAAATG, respectively; F-TNS 4: CTGCAAACTCACCATCCCACA, R-TNS 4: AAGGTGGTGGAGATGGCTTTC, real-time fluorescent quantitative PCR analysis. The expression of BCHE, ADH1A and TNS4 measured by PCR is standardized by the result numerical value of an internal reference gene GAPDH, then Log2 conversion is carried out, bat characteristic score is calculated, the bat characteristic score is further divided into a high group and a low group, the relapse-free survival time of the high group and the low group of patients is analyzed and statistically compared according to the clinical follow-up information of the patients, a survival curve is drawn by Kaplan-Meier survival analysis, and the Log-rank is adopted to test whether the difference has statistical significance so as to reflect the response condition of postoperative auxiliary chemotherapy reaction;

(5) immunohistochemical staining of tumor tissue immune infiltration cell subsets

Paraffinized pancreatic cancer tissue sections were immersed in 3% H₂O₂To quench endogenous peroxidase activity, and then incubated overnight (antibody dilution ratio 1: 100) with anti-CD 4, CD8, CD11C, MPO, CD68 antibodies at 4 ℃ overnight. The next day the sections were incubated with horseradish peroxidase secondary antibody and diaminobenzidine solution. Cell nuclei were stained with harris hematoxylin stain. Subsequently, the level of immune cell infiltration was counted as the number of positive cells per unit area by calculating the interstitial area (mm) on the tissue section²) And the number of different positive immune cells in the region, using the following formula to calculate: total immune cell count/interstitial area (mm)²) (ii) a The optimal cut-off value for each immune subpopulation was calculated according to the john's index of the ROC curve, with high infiltrates defined above the cut-off value and low infiltrates defined below the cut-off value.

(6) Relation between bat characteristic score and pancreatic cancer infiltration immune subgroup

Using CIBERSORT algorithm (https:// CIBERSORT. stanford. edu/index. php), TCGA case chemotherapy sensitive group and drug resistant group were analyzed preliminarily for differences in tumor infiltrating immune cell subpopulations (naive B cells, memory B cells, plasma cells, CD8+ T cells, naive CD4+ T cells, resting memory T cells, activated memory T cells, follicular helper T cells, regulatory T cells, γ δ + T cells, resting NK T cells, activated NK T cells, monocytes, M0 macrophages, M1 macrophages, M2 macrophages, resting dendritic cells, activated dendritic cells, resting mast cells, activated mast cells, eosinophils, neutrophils). And further combining the high-grade group and the low-grade group with the characteristic score of the bat, and analyzing the correlation between the characteristic score of the bat and the infiltration of tumor immune subgroup cells by using independent sample T test or rank sum test and tumor infiltration CD4+ T cells, CD8+ T cells, dendritic cells, macrophages and neutrophils.

(7) Construction and evaluation of Nomogram model for chemotherapy recurrence correlation prediction after pancreatic cancer operation

Postoperative RFS independent prediction factors of pancreatic cancer patients are identified through multi-factor Cox proportional risk regression analysis, the factors are incorporated into a prediction model for Cox regression to calculate partial regression coefficients of variables, and an rms package in R language software is used for establishing graphical Nomogram. And (3) corresponding survival probability is the probability scale axis value corresponding to the position of the bat feature score in the corresponding Nomogram model.

3. The experimental results are as follows:

(1) by obtaining pancreatic cancer cases meeting the inclusion condition, 26 chemotherapy resistant medicine cases and 24 chemotherapy sensitive groups from a TCGA public database. Volcanic plot analysis showed that, in the chemosensitive and resistant groups, a total of 569 DEG were identified, 490 of which were up-regulated and 79 down-regulated (see fig. 1A); the hierarchical cluster analysis plots exhibited the first 50 DEG expression patterns of the chemotherapy sensitive and resistant groups to differentiate the extent of response to postoperative adjuvant chemotherapy (fig. 1B). Using the DEGs obtained above, a one-way Cox survival analysis was performed, and the results showed that a total of 25 DEGs were identified as having significant association with OS (P <0.05, table 1). Next, by Lasso-pendized Cox analysis, it was shown that Butyrylcholinesterase (BCHE), WAP 4-disulfide core domain 3(WFDC3), sulfotransferase family 1B member 1(SULT1B1), India hedgehog signaling molecule (IHH), alcohol dehydrogenase1A (ADH1A), tensin4(TNS4) and charged polypodopsin 4B pseudogene 1(CHMP4BP1) had better identifying potency (FIGS. 2A-2B). Finally, the results of multifactorial Cox survival analyses show that BCHE, ADH1A, and TNS4 are independent prognostic predictors of response to postoperative chemotherapy, BCHE and ADH1A are expressed at levels significantly higher in the chemotherapy-sensitive group than in the chemotherapy-resistant group, while TNS4 is lower than in the chemotherapy-resistant group (P <0.05, fig. 2D-2F). Furthermore, the three genes (BCHE, ADH1A and TNS4) are used to form a new feature score, and the initial letter of the new feature score is taken to form an English word "BAT", so that the new feature score is named as "BAT" feature score. According to the regression coefficient of the multi-factor Cox survival analysis, the 'bat' feature score calculation formula is finally determined as follows: the results show that the characteristic score of the bat in the chemotherapy resistant drug is obviously higher than that of the chemotherapy sensitive group (P <0.05, figure 2C)

TABLE 1 Single factor Cox survival analysis of differential genes associated with overall survival

(2) The "bat" feature scores were calculated from the TCGA data, and the cases were divided into two groups, a low-score group and a high-score group, and Kaplan-Meier survival analysis results showed that the median OS of the patients in the low-score group was 38.4 months (95% confidence interval: 17.3-52.3), significantly higher than 17.5 months (95% confidence interval: 8.6-26.3, P ═ 0.003, fig. 3A) in the high-score group, the 1-year, 2-year and 3-year OS rates were 86.6%, 72.7% and 36.4%, respectively, in the low-score group, and the high-score group was 67.4%, 27.6% and 10.4%. The results of the subject's working curve analysis showed that the area under the curve where the "bat" feature score predicted the level of postoperative adjuvant chemotherapy response and overall survival was 0.734 and 0.780, respectively (fig. 3B-3C). In combination with TCGA clinical data, in single-factor Cox survival analysis, the older (P ═ 0.137), the higher BAT score (P ═ 0.005), and the better ACT response (P ═ 0.037) are important potential risk factors; whereas in the multifactor Cox survival analysis, only the "bat" feature score (P0.040, risk ratio 3.299; 95% confidence interval: 1.058-10.289) remains an independent predictor of OS (table 2).

TABLE 2 independent risk factors for Single and Multi-factor Cox survival analysis affecting overall survival

(3) The results of real-time fluorescent quantitative PCR on 29 included pancreatic cancer tissue specimens showed that the median Δ CT values for BCHE, ADH1A, and TNS4 were 9.1[ interquartile range (IQR): 7.1-10.5) ], 8.8 (IQR: 7.7-10.4), 6.4 (IQR: 4.3-8.3). The results of single-factor and multi-factor Cox survival analyses further demonstrate that the "bat" feature score is an independent risk factor affecting RFS in pancreatic cancer patients receiving adjuvant chemotherapy post-operatively (P0.014, risk ratio 5.172, 95% confidence interval: 1.403-19.063, table 3). The median RFS of the patients with high bat feature score is 9 months (95% confidence interval: 6.0-12.0), is obviously shorter than that of the patients with low bat feature score (not reached, figure 3D), the survival difference between the two groups is obvious and has statistical significance, and the result prompts that the bat feature score can effectively reflect the postoperative adjuvant chemotherapy effect of pancreatic cancer.

TABLE 3 Single and multifactor Cox survival analysis of the associated independent risk factors affecting postoperative adjuvant chemotherapy recurrence of pancreatic cancer

(4) The results of CIBERSORT analysis of TCGA database data showed a clear upward trend for activated CD4+ T cells, CD8+ T cells, resting dendritic cells, M0 macrophages (fig. 4). To further verify, fig. 5 is the results of immunohistochemical staining of tissue chips from two pancreatic cancer cases, showing the infiltration of CD4+ T cells, CD8+ T cells, CD11C + dendritic cells, CD68+ macrophages, and MPO + neutrophils in the cancer tissues, respectively. Case 1 represents a high infiltration of CD4+ T cells, a high infiltration of CD8+ T cells, a moderate infiltration of CD11C + dendritic cells, a high infiltration of CD68+ macrophages, and a low infiltration of MPO + neutrophils; while case 2 represented low infiltration of CD4+ T cells, low infiltration of CD8+ T cells, moderate infiltration of CD11C + dendritic cells, moderate infiltration of CD68+ macrophages, and high infiltration of MPO + neutrophils, with RFS significantly longer for case 1 than for case 2. Statistical analysis found that tumor infiltration CD4+ and CD8+ T cells were significantly reduced in tissues with higher "bat" feature scores (P values of 0.015 and 0.021, respectively, fig. 6A-6B); however, there were no significant differences in macrophage, neutrophil, and dendritic cell infiltration (fig. 6C-6E). In addition, it was also found that patients with a high "bat" feature score were more prone to receive post-operative chemotherapy based on gemcitabine regimens (P ═ 0.022), and that a higher "bat" feature score is generally predictive of early relapse, which may indicate that patients with a high "bat" feature score are more susceptible to gemcitabine resistance (table 4).

TABLE 4 correlation analysis between "Bat" feature score and pancreatic cancer clinical features and immune-infiltrating cell subpopulation

(5) And (3) incorporating the characteristic score of the bat into corresponding Cox regression to construct a Nomogram model for predicting RFS 6 months, 12 months and 18 months after postoperative adjuvant chemotherapy, i.e. incorporating indexes with statistical significance in recurrence-free survival time multi-factor Cox proportional risk regression analysis into corresponding Cox regression to construct the Nomogram model. RFS survival probability of months 6, 12 and 18 (as shown in FIG. 7) is determined by scaling the axial value of the survival probability of months 6, 12 and 18 corresponding to the position of the characteristic score of "bat" (as shown in FIG. 7)

While the invention has been described with respect to a preferred embodiment, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention. Those skilled in the art can make various changes, modifications and equivalent arrangements, which are equivalent to the embodiments of the present invention, without departing from the spirit and scope of the present invention, and which may be made by utilizing the techniques disclosed above; meanwhile, any changes, modifications and variations of the above-described embodiments, which are equivalent to those of the technical spirit of the present invention, are within the scope of the technical solution of the present invention.