Lung adenocarcinoma prognosis marker and model

Abstract

The invention provides a lung adenocarcinoma prognostic marker and a model, constructs a lung adenocarcinoma prognostic model based on mRNA and lncRNA related to programmed cell death (PCD), and synthesizes bioinformatics analysis and experiments to explore the biological effect of the lncRNA model in lung adenocarcinoma. The prognosis model covers all PCD pathways and various molecules, has stable prognosis prediction accuracy, and can provide a potential molecular target for personalized treatment. According to the method disclosed by the invention, the PCD-related mRNA and lncRNA are integrated to construct a lung adenocarcinoma diagnosis model suitable for Chinese population, so that the identification capability of tumor patient classification is enhanced from multiple aspects, the accuracy and reliability of prognosis prediction are improved, powerful support can be provided for promoting personalized treatment strategies of lung adenocarcinoma patients, and the method is suitable for popularization and application.

Description

Technical Field

[0001] The present invention relates to the technical field of tumor molecular biology, and in particular to a lung adenocarcinoma prognostic marker and model. Background Art

[0002] Lung adenocarcinoma is the most prevalent subtype of lung cancer, accounting for approximately 40-50% of lung cancer cases. However, due to treatment heterogeneity and the emergence of multidrug resistance, the prognosis of lung adenocarcinoma is dismal. Currently, prognostic assessment of lung adenocarcinoma is primarily based on clinicopathological features. With the advancement of high-throughput sequencing technology, cancer prognostic models based on genes associated with biological pathways have effectively improved the biological interpretability and predictive accuracy of these models. Programmed cell death (PCD) is a predetermined cell death process executed in response to internal signals. Different PCD mechanisms have been shown to play a key role in lung adenocarcinoma, and its dysregulation is a key factor in tumor development and drug resistance. Messenger RNA (mRNA) is widely used as a biomarker for cancer development and progression. Furthermore, long noncoding RNA (lncRNA) possesses a certain degree of specificity and can participate in regulating the PCD pathway, affecting tumor cell growth and death. Constructing a prognostic prediction model by integrating the association between mRNA, lncRNA, and PCD may provide new insights into the pathogenic mechanisms of lung adenocarcinoma, identify high-risk individuals, and provide more effective and targeted treatment options.

[0003] Current studies on constructing prognostic models for various PCD pathways all focus on a single or two cell death pathways. Even if all PCD pathways are considered, the role of lncRNA and the combined effects of lncRNA and mRNA are not taken into account. In addition, the model performance evaluation and predictive efficacy are uneven. It is worth noting that no PCD-related prognostic model has been constructed for the Chinese population. Summary of the Invention

[0004] To address the above problems, the present invention provides a prognostic marker and model for lung adenocarcinoma, constructs a prognostic model based on PCD-related mRNA and lncRNA, and integrates the two to construct a PCD-related transcriptomic score and a comprehensive prognostic model of PCD-related mRNA and lncRNA. The prognostic model constructed by the present invention covers all PCD pathways and multiple molecules, has robust prognostic prediction accuracy, and can provide potential molecular targets for personalized treatment.

[0005] One of the technical solutions of the present invention is to provide a prognostic marker for lung adenocarcinoma, which covers the molecular characteristics of various PCD pathways. These pathways play an important role in maintaining tissue stability, development and disease defense. Their molecular characteristics are of great significance for the prognosis of lung adenocarcinoma. Specifically, they include 26 programmed cell death mRNAs (PCDmRs) and 12 programmed cell death lncRNAs (PCDlncs). Among them, PCDmRs include FLNC, BAK1, GAPDH, SLC2A1, LDHA, ERO1A, NTSR1, ANGPTL4, LYPD3, CDK5R1, VDAC2, SPHK1, SLC7A5, LRP8, AVEN, PTPRH, H2AC19, FOSL1, PGAM4, KNG1, TFAP2A, PPT2, BCL2L10, MOAP1, YKT6, and AKT1S1; PCDlncs include LINC00707, FAM83A-AS1, LINC02323, LINC01116, ITGB1-DT, CYTOR, EMSLR, LINC00968, LINP1, MIR210HG, FENDRR, and OGFRP1.

[0006] The second technical solution of the present invention is to provide a lung adenocarcinoma prognosis model, which is: PTS=I1×PMS+I2×PLS; where I1 and I2 are the regression coefficients of PMS and PLS respectively.

[0007] The PMS is a random survival forest (RSF) model, and the PCD-related mRNA expression score is obtained based on PCD-related mRNA; PCD-related mRNAs are: FLNC, BAK1, GAPDH, SLC2A1, LDHA, ERO1A, NTSR1, ANGPTL4, LYPD3, CDK5R1, VDAC2, SPHK1, SLC7A5, LRP8, AVEN, PTPRH, H2AC19, FOSL1, PGAM4, KNG1, TFAP2A, PPT2, BCL2L10, MOAP1, YKT6, AKT1S1; The PLS was a random survival forest (RSF) model, and the PCD-related lncRNA expression scores were obtained based on PCD-related lncRNAs; PCD-related lncRNAs were: LINC00707, FAM83A-AS1, LINC02323, LINC01116, ITGB1-DT, CYTOR, EMSLR, LINC00968, LINP1, MIR210HG, FENDRR, and OGFRP1.

[0008] Furthermore, the prognostic model was constructed by the following method: (1) Collect PCD-related mRNAs from databases and literature as PCDmR candidate genes, and collect PCD-related lncRNAs from databases and literature as PCDlnc candidate genes; (2) Based on PCDmR candidate genes, models were constructed in the discovery cohort and tested in the validation cohort. The C-index of each model in the discovery cohort and validation cohort was calculated. The model with the highest average C-index was considered the optimal model PMS, which was used to obtain PCD-related mRNA expression scores based on PCD-related mRNAs. PCD-related mRNAs were: FLNC, BAK1, GAPDH, SLC2A1, LDHA, ERO1A, NTSR1, ANGPTL4, LYPD3, CDK5R1, VDAC2, SPHK1, SLC7A5, LRP8, AVEN, PTPRH, H2AC19, FOSL1, PGAM4, KNG1, TFAP2A, PPT2, BCL2L10, MOAP1, YKT6, AKT1S1; (3) Based on PCD lnc candidate genes, models were constructed in the discovery cohort and tested in the validation cohort. The C-index of each model in the discovery cohort and validation cohort was calculated, and the model with the highest average C-index was considered the optimal model. PLS was used to obtain PCD-related lncRNA expression scores based on PCD-related lncRNAs; PCD-related lncRNAs were: LINC00707, FAM83A-AS1, LINC02323, LINC01116, ITGB1-DT, CYTOR, EMSLR, LINC00968, LINP1, MIR210HG, FENDRR, and OGFRP1; (4) A prognostic model was constructed using a multivariate Cox regression model: PTS=I1×PMS+I2×PLS; where I1 and I2 are the regression coefficients of PMS and PLS respectively.

[0009] The third technical solution of the present invention is to provide a reagent for detecting the expression level of the above-mentioned markers for use in preparing a product for evaluating the prognosis of lung adenocarcinoma and / or predicting the efficacy of immunotherapy for lung adenocarcinoma.

[0010] Furthermore, the above reagents include substances for detecting the expression level of combined genes in the markers by sequencing technology, nucleic acid hybridization technology, and nucleic acid amplification technology.

[0011] The fourth technical solution of the present invention is to provide the above-mentioned detection reagent for detecting the expression level of the marker.

[0012] The fifth technical solution of the present invention is to provide a kit comprising the above-mentioned detection reagent.

[0013] The sixth technical solution of the present invention is to provide a system for evaluating the prognosis of lung adenocarcinoma and / or predicting the efficacy of immunotherapy for lung adenocarcinoma, wherein the system includes a calculation unit, which calculates a score using the prognostic model.

[0014] The benefits of the present invention lie in that a lung adenocarcinoma diagnostic model suitable for the Chinese population is constructed by integrating PCD-related mRNA and lncRNA, which enhances the ability to classify tumor patients from multiple levels, improves the accuracy and reliability of prognosis prediction, and can provide strong support for promoting personalized treatment strategies for lung adenocarcinoma patients. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 To validate the efficacy of the PTS prognostic model.

[0016] Figure 2 Figure 3 is the time-dependent ROC curve of the PTS prognostic model. DETAILED DESCRIPTION

[0017] The following examples are used to further illustrate the present invention. Their purpose is to illustrate the present invention and should not be construed as limiting the scope of the present invention. Unless otherwise specified, all references are by weight and weight percentage.

[0018] Unless otherwise specified, the raw materials used in the present invention are conventional commercial products; the methods used in the present invention are conventional methods in the art unless otherwise specified.

[0019] The embodiments of the present invention are further described below with reference to a number of embodiments.

[0020] It should be clear that the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0021] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments and are not intended to limit the present application. The singular forms "a", "an", "the" and "the" used in the embodiments of the present application and the appended claims are also intended to include plural forms unless the context clearly indicates otherwise.

[0022] Example 1 (1) Patients with lung adenocarcinoma from the Second Affiliated Medical School of Zhejiang University School of Medicine (SAHZU-LUAD), The Cancer Genome Atlas (TCGA), and GSE72094 cohort were selected. A two-stage design was used, with the TCGA-LUAD cohort as the discovery cohort and SAHZU-LUAD and GSE72094 as validation cohorts. Both the discovery cohort and validation cohort included RNA-seq sequencing data from multiple lung adenocarcinoma lesion tissue sections and adjacent tissue sections.

[0023] (2) Collect PCD-related mRNAs from databases and literature as PCDmR candidate genes, and collect PCD-related lncRNAs from databases and literature as PCDlnc candidate genes; (3) Construction and validation of a prognostic model based on PCD-related mRNA Univariate Cox regression analysis identified 243 stable prognostic PCD mRNAs. For candidate genes, 41 combinations of nine machine learning algorithms were used to construct models in the discovery cohort, and these 41 models were tested in the validation cohort. For each model, its C-index across all datasets was calculated, and the model with the highest average C-index was considered the optimal model. This led to the PCD-related mRNA transcriptional score (PMS) constructed based on PCD-related mRNAs.

[0024] The nine machine learning algorithms used in this study include random survival forest (RSF), elastic network (Enet), least absolute shrinkage and selection operator regression (LASSO), ridge regression, CoxBoost, partial least squares regression for Cox (plsRcox), supervised principal component analysis (SPCA), generalized boosted regression modeling (GBM), and survival support vector machine (SSVM). LASSO, CoxBoost, Enet, and RSF can be used for feature selection. The following table lists the R packages and functions used by each machine learning algorithm.

[0025] algorithm R packages function Elastic Net (Enet) glmnet (version 4.1-7) cv.glmnet Least Absolute Shrinkage and Selection Operator Regression (LASSO) glmnet (version 4.1-7) cv.glmnet Ridge Regression glmnet (version 4.1-7) cv.glmnet Random Survival Forest (RSF) randomForestSRC (version 3.2.3) rfsrc Cox Boosting Algorithm (CoxBoost) CoxBoost (version 1.4) CoxBoost Partial Least Squares Regression for Cox (plsRox) plsRcox (version 1.7.7) plsRcox Supervised Principal Component Analysis (SPCA) superpc (version 1.12) super.cv Generalized Boosted Regression Modeling (GBM) gbm (version 2.1.8) gbm Survival Support Vector Machine (SSVM) survivalsvm (version 0.0.5) survivalsvm The optimal model building method was selected as RSF, with an average C-index of 0.80.

[0026] RSF is a random forest-based survival analysis method that handles right-censored data (i.e., survival time data) by constructing multiple decision trees. The survival trees are then used to predict individual survival functions and risk scores. The model has two parameters: ntree (the number of trees in the forest) and mtry (the number of variables randomly selected at each split). A grid search using cross-validation was performed on ntree and mtry to identify the combination with the best concordance index (C-index) value. This resulted in the identification of 26 PCDmRs associated with stable prognosis: FLNC, BAK1, GAPDH, SLC2A1, LDHA, ERO1A, NTSR1, ANGPTL4, LYPD3, CDK5R1, VDAC2, SPHK1, SLC7A5, LRP8, AVEN, PTPRH, H2AC19, FOSL1, PGAM4, KNG1, TFAP2A, PPT2, BCL2L10, MOAP1, YKT6, and AKT1S1. Among them, 19 were related to apoptosis, 9 to ADCD, and 5 to ferroptosis. In addition, ICD, necroptosis, and oxeiptosis each had 2 related genes, while disulfide death, LDCD, MPTDN, and pyroptosis each had 1 related gene.

[0027] (4) Construction and validation of a prognostic model based on PCD-related lncRNA Univariate Cox regression analysis identified 125 stable, prognostic PCD lncRNAs. Forty-one combinations of nine machine learning algorithms were used to construct models for the candidate genes in the discovery cohort, and these 41 models were tested in the validation cohort. For each model, its C-index across all datasets was calculated, and the model with the highest average C-index was considered the optimal model. This led to the PCD-related lncRNA transcriptional score (PLS) constructed based on the PCD-related lncRNAs.

[0028] The optimal PLS model construction method was also RSF, with an average C-index of 0.80. Twelve PCDlncs associated with stable prognosis were screened: LINC00707, FAM83A-AS1, LINC02323, LINC01116, ITGB1-DT, CYTOR, EMSLR, LINC00968, LINP1, MIR210HG, FENDRR, and OGFRP1.

[0029] (5) Construction and validation of PCD-related transcriptomic expression scores The calculation formula of PTS was obtained by incorporating PMS and PLS into the multivariate Cox regression model: PTS = -0.065×PMS+0.160×PLS.

[0030] Example 2 Evaluation of the predictive efficacy of expression scoring models (1) The PMS and PLS scoring models were applied to two independent validation cohorts for validation. To further evaluate the predictive efficacy of the prognostic model, the expression scoring model obtained in the previous step was applied to two independent validation cohorts for validation. For each subject in the validation cohort, the score was calculated and grouped according to the established scoring model formula. At the same time, considering the differences between the gene sequencing technologies of each cohort, some key lncRNAs identified in the TCGA-LUAD cohort and SAHZU-LUAD cohort were not described in the GSE72094 validation cohort. Therefore, this study used the alternative lncRNA method for cross-platform validation. The selected alternative lncRNA was significantly correlated with the initial lncRNA (FDR < 0.05) and had the largest Pearson correlation coefficient.

[0031] (2) In the discovery cohort and validation cohort, multivariate Cox regression analysis was used to explore the effectiveness of expression scores and other clinical characteristics (including age, clinical stage, smoking history, and gender) as independent prognostic factors. The effect of one or more variables on patient prognosis was evaluated after controlling for the influence of other relevant variables. If the variable P If the value is still significant, the variable can be considered to have an independent prognostic effect.

[0032] (3) Kaplan-Meier (KM) survival analysis was used to evaluate the survival probability of the subjects in the high and low score groups in the discovery and validation sets, and survival curves were drawn. The survival curves were compared using the log-rank test to test whether the difference in survival between the two groups was statistically significant. The HR and its 95% confidence interval (CI) were calculated to assess the relative magnitude of the survival risk between the different score groups. The above analyses were performed using the R package survival (version 3.5-5) and the R package survmier (version 0.4.9).

[0033] To quantitatively evaluate the predictive performance of the prognostic model, the R package timeROC (version 0.4) was used to draw the time-dependent receiver operating characteristic (ROC) curve, and the area under the curve (AUC) at different time points was calculated to measure the model's predictive performance. A higher AUC value indicates that the model has better classification performance and can more accurately distinguish patient groups with different prognoses. The receiver operating characteristic (ROC) curve and the area under the curve (AUC) measure the model's predictive performance. Figure 1 As shown in the figure, the risk score of patients in the high-risk group increased significantly over time and had a higher mortality rate, while the risk score of patients in the low-risk group was relatively stable and had a higher survival rate, indicating that the model has certain effectiveness in predicting patient prognosis and can be used for further clinical evaluation and decision-making.

[0034] Example 3 140 pairs of lung adenocarcinoma and adjacent tissue samples from SANZU were selected, and quantitative reverse transcription polymerase chain reaction (qRT-PCR) was used to experimentally verify the expression levels of key lncRNAs in lung adenocarcinoma and adjacent tissues.

[0035] 1. Total RNA Extraction (1) Tissue grinding and lysis (1.1) Prepare Buffer RLT Plus by adding β-mercaptoethanol at a ratio of 100:1. Grind the tissue in liquid nitrogen and transfer it to Buffer PLT Plus (β-ME). Mix thoroughly by pipetting. For tissue <10 mg, add 350 μL of Buffer RLT Plus; for tissue <30 mg, add 600 μL of Buffer RLT Plus.

[0036] (1.2) Transfer the tissue grinding solution to a DNA spin column, place it in a 2 mL collection tube, and centrifuge at 17,000 × g for 45 seconds.

[0037] (2) Digestion of proteins bound to nucleic acids (2.1) Add 80 μL of Proteinase K to each tissue collection tube and pipette to mix thoroughly.

[0038] (2.2) Add 350 μL of 96-100% ethanol and mix thoroughly by pipetting. Let stand at room temperature for 10 minutes. (2.3) Add 750 μL of 96-100% ethanol and mix by pipetting.

[0039] (3) Binding to RNA Place the RNeasy spin column in a 2 mL collection tube and centrifuge at 17,000 × g for 30 seconds, with 600 μL of sample (including the precipitate) being transferred until the column is complete.

[0040] (4) Purification (4.1) Centrifuge 500 μL Buffer RPE at 17,000 × g for 30 seconds and discard the waste liquid.

[0041] (4.2) Prepare the DNase I incubation mix: 7:1 Buffer RDD + DNase I stoeksolution and mix thoroughly by gently inverting.

[0042] (4.3) Centrifuge 500 μL Buffer FRN at 17,000 × g for 30 seconds, retain the liquid after centrifugation, and discard the waste liquid.

[0043] (4.4) Centrifuge 500 μL Buffer RPE at 17,000 × g for 30 seconds and discard the waste liquid.

[0044] (4.5) Add 500 μL of 96-100% ethanol and centrifuge at 17,000 × g for 2 minutes. Discard the waste liquid.

[0045] (5) Dry and wash (5.1) Place the RNeasy spin column in 1.5 mL Safe-Lock tubes and allow to air dry for 2 minutes.

[0046] (5.2) Add 30 μL RNase-free water, let stand at room temperature for 5 minutes, and centrifuge at 8000 × g for 1 minute.

[0047] (5.3) RNA concentration was measured using a Qubit Flex Fluorometer. RNA integrity (RIN) was assessed using a TapeStation 4150. Samples with a RIN ≥ 7.0 were used for subsequent testing.

[0048] 2. PCR Primer Design Primer sequences for each gene were designed using the website Primer-Blast. The primer sequences for this study are as follows:

[0049] 3. qRT-PCR Detection Using Fluidigm GE48.48 IFC Chip (1) Reverse transcription of cDNA The extracted total RNA (≤1 μg) was used according to the instructions of the HiScript III RT SuperMix for qPCR reverse transcription kit and incubated in a PCR instrument to obtain a 20 μL cDNA sample.

[0050] (2) Sample pre-amplification (2.1) Premixed primers (2.1.1) Thaw the primers (primer concentration is 10 μM), vortex, and centrifuge briefly before use; (2.1.2) Prepare the Pooled Delta Gene assay mix in a 1.5 mL centrifuge tube (final concentration of upstream and downstream primers: 500 nM). Add 1 μL of each primer pair and make up to 200 μL with DNA suspension buffer.

[0051] Vortex to mix, centrifuge briefly and set aside.

[0052] (2.2) Prepare the pre-amplification reaction system (2.2.1) Thaw the Preamp Master Mix, vortex, centrifuge briefly, and store on ice until ready to use. (2.2.2) Prepare the preamplification pre-mix in a 1.5 mL centrifuge tube. The volume for a single reaction should include: 1 μL Preamp Master Mix, 0.5 μL Pooled Delta Gene assay mix, and 2.25 μL DNase-free water. Calculate the volume of the preamplification pre-mix based on the desired sample volume. Vortex to mix, centrifuge briefly, and store on ice.

[0053] (2.3) Samples are mixed with the pre-amplification system In a 96-well plate, add 1.25 μL of cDNA sample and 3.75 μL of preamplification pre-mix to each well, labeled Sample Plate. Vortex for 5 seconds and centrifuge at 1000 × g for 1 minute.

[0054] (2.4) Pre-amplification reaction Use the Preamp program in a PCR instrument and a temperature cycle program to perform PCR amplification. After the reaction is complete, centrifuge at 1000 × g for 1 minute. The reaction conditions are as follows.

[0055]

[0056] Exonuclease I digestion to remove primers from the pre-amplification reaction (2.4.1) Prepare Diluted Exo I in a 1.5 mL centrifuge tube. The volume of a single reaction system includes 1.4 μL DNase-free water, 0.2 μL Exonuclease I Reaction Buffer, and 0.4 μL Exonuclease I. Vortex to mix, centrifuge briefly, and store on ice until ready to use.

[0057] (2.4.2) Add the prepared Diluted ExoI to the reaction product from the previous step, 2 μL per well, to obtain the enzyme digestion reaction system. Vortex for 5 seconds and centrifuge at 1000 × g for 1 minute.

[0058] (2.4.3) Perform the PCR digestion procedure using the ExoI program in a PCR instrument: digest at 37°C for 30 minutes, then incubate at 80°C for 15 minutes to terminate the reaction. After the reaction is complete, centrifuge at 1000 × g for 1 minute.

[0059] (2.5) Product dilution Adjust the dilution factor based on the Ct value. In this experiment, a 5-fold dilution was used: add 18 μL of DNA suspension buffer to the reaction product from the previous step, for a final volume of 25 μL. Vortex for 5 seconds and centrifuge at 1000 × g for 1 minute.

[0060] (3) qRT-PCR reaction and detection using Fluidigm GE 48.48 IFC chip (3.1) Prepare the assay mix Prepare the assay mix in a 96-well plate. Add primers according to the assay map. Each inlet should contain: 2.5 μL of 2× Assay Loading Reagent, 2 μL of DNA Suspension Buffer, 0.25 μL of 100 μM upstream primer, and 0.25 μL of 100 μM downstream primer. Vortex for 20 seconds and centrifuge at 1000 × g for 1 minute.

[0061] (3.2) Prepare the Sample Mix (3.2.1) Thaw the reagents, vortex, and centrifuge briefly. Keep on ice until ready to use. Prepare the Sample Pre-mix in a 1.5 mL centrifuge tube. Each inlet should contain 2.5 μL of 2× SsoFast EvaGreen Supermix with lowROX and 0.25 μL of 20× DNA Binding Dye.

[0062] (3.2.2) Prepare the sample mix. Each inlet should contain 2.75 μL of sample pre-mix and 2.25 μL of pre-amplification product sample plate. Vortex for 20 seconds and centrifuge at 1000 × g for 1 minute.

[0063] (3.3) Add 300 μL of control line fluid to the IFC chip through the cross-shaped injection holes on both sides. Remove the thermal insulation film from the bottom of the IFC and place it on the Fluidigm Juno for priming.

[0064] (3.4) After the Prime run, add the Assay Mix to the left well of the IFC chip and the Sample Mix to the right well of the IFC chip, 5 μL per well. Load the samples using the Fluidigm Juno. After loading, perform the qRT-PCR reaction and signal acquisition within 1 hour.

[0065] (3.5) Place the loaded IFC chip in the Fluidigm BioMark HD and select the RT-PCR Data Collection program, GE Fast 48x48 PCR + Meltv2.pcl, for the reaction and signal acquisition. Review the data using SBIReal-Time PCR software.

[0066] (4) Calculation of relative expression levels of qRT-PCR target genes In this study, two genes, GAPDH and ACTB, were used as reference genes to improve the accuracy and reliability of the results. The calculation method used in the studies of Hellemans et al. and Vandesompele et al. (86, 87) was used. The formula for calculating the relative expression of the target genes is as follows:

[0067] Where, GOI is the target gene, REF is the reference gene, and E is the primer efficiency.

[0068] First, the amplification efficiency of all primers was verified by preliminary experiments. Therefore, it is assumed here that the amplification efficiency of all primers is 100%, and the calculated primer efficiency E = 2. Then, the average value of the Ct value of the control group was selected as the Ct value of the calibration sample (Calibrator Ct), and the difference between the target gene Ct value of each sample and the calibration sample Ct value was calculated to obtain the ΔCt value. The relative quantity (RQ) value of the target gene of each sample was calculated according to the formula. Subsequently, the geometric mean of the RQ value of the reference gene was calculated, and the relative gene expression value was obtained by comparing the RQ value of the target gene with the ratio of the geometric mean RQ value of the reference gene.

[0069] Subsequently, paired t-tests were used to compare gene expression differences between tumor tissues and adjacent adjacent tissues, and to further analyze gene expression differences in tumor tissues of different clinical stages. The results showed that compared with adjacent adjacent tissues, LINC00707, FAM83A-AS1, LINC02323, CYTOR, EMSLR, LINP1, MIR20HG, and OGFRP1 had significantly higher expression levels in lung adenocarcinoma tissues, while LINC00968 and FENDRR expression were relatively low in lung adenocarcinoma tissues, with P values ​​all less than 0.05. The differences in gene expression in tumor tissues according to different clinical stages showed that LINC02323 expression was significantly increased in stage III tissues compared with stage I tissues ( P <0.05). This lung adenocarcinoma prognostic marker has a high prediction accuracy.

[0070] The above embodiments describe in detail the structure, features and effects of the present invention. The above are only preferred embodiments of the present invention. Any changes made in accordance with the concept of the present invention, or modifications to equivalent embodiments with equivalent changes, which do not exceed the scope covered by the specification, should be within the scope of protection of the present invention.

Claims

1. A prognostic marker for lung adenocarcinoma, characterized in that: It includes 26 programmed cell death mRNAs (PCDmRs) and 12 programmed cell death lncRNAs (PCDlncs); among them, PCDmRs include FLNC, BAK1, GAPDH, SLC2A1, LDHA, ERO1A, NTSR1, ANGPTL4, LYPD3, CDK5R1, VDAC2, SPHK1, SLC7A5, LRP8, AVEN, PTPRH, H2AC19, FOSL1, PGAM4, KNG1, TFAP2A, PPT2, BCL2L10, MOAP1, YKT6, and AKT1S1; PCDlncs include LINC00707, FAM83A-AS1, LINC02323, LINC01116, ITGB1-DT, CYTOR, EMSLR, LINC00968, LINP1, MIR210HG, FENDRR, and OGFRP1.

2. A lung adenocarcinoma prognosis model, characterized in that: The prognostic model is: PTS=I1×PMS+I2×PLS; where I1 and I2 are the regression coefficients of PMS and PLS respectively; The PMS is a random survival forest (RSF) model, and the PCD-related mRNA expression score is obtained based on PCD-related mRNA; PCD-related mRNAs are: FLNC, BAK1, GAPDH, SLC2A1, LDHA, ERO1A, NTSR1, ANGPTL4, LYPD3, CDK5R1, VDAC2, SPHK1, SLC7A5, LRP8, AVEN, PTPRH, H2AC19, FOSL1, PGAM4, KNG1, TFAP2A, PPT2, BCL2L10, MOAP1, YKT6, AKT1S1; The PLS was a random survival forest (RSF) model, and the PCD-related lncRNA expression scores were obtained based on PCD-related lncRNAs; PCD-related lncRNAs were: LINC00707, FAM83A-AS1, LINC02323, LINC01116, ITGB1-DT, CYTOR, EMSLR, LINC00968, LINP1, MIR210HG, FENDRR, and OGFRP1.

3. The prognostic model according to claim 2, characterized in that It is constructed by the following method: (1) Collect PCD-related mRNAs from databases and literature as PCDmR candidate genes, and collect PCD-related lncRNAs from databases and literature as PCDlnc candidate genes; (2) Based on PCDmR candidate genes, models were constructed in the discovery cohort and tested in the validation cohort. The C-index of each model in the discovery cohort and validation cohort was calculated. The model with the highest average C-index was considered the optimal model PMS, which was used to obtain PCD-related mRNA expression scores based on PCD-related mRNAs. PCD-related mRNAs were: FLNC, BAK1, GAPDH, SLC2A1, LDHA, ERO1A, NTSR1, ANGPTL4, LYPD3, CDK5R1, VDAC2, SPHK1, SLC7A5, LRP8, AVEN, PTPRH, H2AC19, FOSL1, PGAM4, KNG1, TFAP2A, PPT2, BCL2L10, MOAP1, YKT6, AKT1S1; (3) Based on PCD lnc candidate genes, models were constructed in the discovery cohort and tested in the validation cohort. The C-index of each model in the discovery cohort and validation cohort was calculated, and the model with the highest average C-index was considered the optimal model. PLS was used to obtain PCD-related lncRNA expression scores based on PCD-related lncRNAs; PCD-related lncRNAs were: LINC00707, FAM83A-AS1, LINC02323, LINC01116, ITGB1-DT, CYTOR, EMSLR, LINC00968, LINP1, MIR210HG, FENDRR, and OGFRP1; (4) A prognostic model was constructed using a multivariate Cox regression model: PTS=I1×PMS+I2×PLS; where I1 and I2 are the regression coefficients of PMS and PLS respectively.

4. Use of a reagent for detecting the expression level of the marker according to claim 1 in the preparation of a product for evaluating the prognosis of lung adenocarcinoma and / or predicting the efficacy of immunotherapy for lung adenocarcinoma.

5. The application according to claim 4, characterized in that: The reagents include substances for detecting the expression levels of combined genes in markers by sequencing technology, nucleic acid hybridization technology, and nucleic acid amplification technology.

6. A detection reagent, characterized in that The detection reagent is used to detect the expression level of the marker according to claim 1.

7. A kit, characterized in that The kit comprises the detection reagent according to claim 6.

8. A system for evaluating the prognosis of lung adenocarcinoma and / or predicting the efficacy of immunotherapy for lung adenocarcinoma, characterized in that: The system comprises a calculation unit for calculating a score using the prognostic model according to claim 2 .

Images