Early lung adenocarcinoma air cavity dissemination related prognosis model and construction method and program product thereof
By constructing a prognostic model related to airspace dissemination of early-stage lung adenocarcinoma, using the LASSO-Cox regression model to screen marker genes, and establishing a risk scoring model, the problem of high recurrence risk in patients with early-stage lung adenocarcinoma was solved, the accuracy and universality of survival rate prediction were improved, and better support for treatment decisions was provided.
Patent Information
- Application Number
- CN202510781904.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-06-12
- Publication Date
- 2025-09-16
AI Technical Summary
The existing technology lacks an effective prognostic model related to airspace dissemination in early-stage lung adenocarcinoma, which leads to a high risk of recurrence or metastasis in patients with early-stage lung adenocarcinoma, affecting the patient's survival rate and prognosis.
A prognostic model related to airspace dissemination of early-stage lung adenocarcinoma was constructed. By obtaining the patient's RNA expression profile data and clinical data, the LASSO-Cox regression model was used to screen marker genes, and a risk scoring model was established. The patients were divided into low-risk and high-risk groups to provide accurate predicted survival assessment.
The seven-gene model has improved the accuracy of predicting survival rates for patients with early-stage lung adenocarcinoma. Validated by multiple data sets, it has shown that the C-index of the seven-gene model is higher, significantly superior to single genes or other gene combinations, and provides more accurate guidance for treatment decisions.
Smart Images
Figure CN120656559A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of tumor molecular biology, and in particular to a prognostic model associated with airspace dissemination of early-stage lung adenocarcinoma, a construction method thereof, and a program product. Background Art
[0002] Lung cancer is the most common cancer worldwide and one of the leading causes of cancer-related death. Non-small cell lung cancer (NSCLC) is the predominant histological subtype of lung cancer, with lung adenocarcinoma (LUAD) being the predominant type. Over the past three decades, the five-year overall survival rate for lung cancer has been less than 20%. Despite surgical treatment, patients with early-stage (pathological stage I) LUAD remain at high risk of recurrence or metastasis. Identifying biomarkers for lung adenocarcinoma, improving understanding of the molecular mechanisms underlying lung adenocarcinoma, and developing novel treatment strategies are essential to improving patient outcomes.
[0003] Airway spread is a unique metastatic pattern in lung tumors, distinct from lymph node and hematogenous metastasis, and refers to the spread of tumor cells through the airways to distant sites. This concept was formally proposed by the World Health Organization in 2015, and numerous studies have reported a correlation between airway spread and poor prognosis in cancer patients. For example, a study by Kadota et al. demonstrated that airway spread (STAS) is a significant pattern of recurrence in early-stage non-small cell lung cancer and significantly impacts recurrence-free survival (RFS), overall survival (OS), and lung cancer-specific survival (LCSS). Specifically, among patients with stage IA non-mucinous lung adenocarcinoma, the 5-year RFS for those with STAS was 91.8% (no STAS), 79.0% (STAS I), and 60.5% (STAS II), while the 5-year OS was 95.2%, 88.3%, and 74.1%, respectively (Kadota K, Nitadori J, Sima CS, et al. Tumor spread through air spaces is an important pattern of invasion and impacts the frequency and location of recurrences after limited resection for small stage Ilung adenocarcinomas. J Thorac Oncol. 2015;10:806–814.). In addition, Uruga et al. further confirmed the prognostic value of airspace dissemination in early-stage lung adenocarcinoma. Through semiquantitative evaluation, they found that the presence of STAS can significantly predict the patient's risk of recurrence (Uruga H, Fujii T, Fujimori S, et al. Semiquantitative assessment of tumor spread through air spaces (STAS) inearly-stage lung adenocarcinomas. J Thorac Oncol. 2017;12:1046–1051.); however, there is no report on the use of gene combinations related to airspace dissemination characteristics for prognosis prediction. Summary of the Invention
[0004] The purpose of the present invention is to provide a prognostic model for airspace dissemination in early-stage lung adenocarcinoma to address the above-mentioned problems. To achieve the above-mentioned objectives, the present invention adopts the following technical solutions: A method for constructing a prognostic model related to airspace dissemination in early-stage lung adenocarcinoma comprises the following steps: S1: Obtain RNA expression profile data and corresponding clinical data of patients with lung adenocarcinoma; S2: Label the airway dissemination characteristics of lung adenocarcinoma patients and divide them into airway dissemination-positive and airway dissemination-negative patients. Based on the labeled airway dissemination characteristics, perform differential analysis on RNA expression profile data to obtain a gene set related to airway dissemination characteristics. S3: Based on the gene set related to airway dissemination characteristics and the corresponding clinical data, a regression model is established to screen out marker genes and obtain the corresponding regression coefficient of each marker gene, thereby forming a prognostic model.
[0005] Preferably, the prognostic model is evaluated by risk scoring, and the risk score is calculated as follows: Risk Score =
[0006] Where n represents the total amount of marker genes; represents the expression level of the i-th marker gene; represents the regression coefficient of the i-th marker gene.
[0007] Preferably, the basis for marking the airspace dissemination characteristics of the lung adenocarcinoma patient is the pathological section of the lung adenocarcinoma patient.
[0008] Preferably, the regression model is a LASSO-Cox regression model.
[0009] Preferably, in the process of establishing the LASSO-Cox regression model, the bootstrap method is used to perform penalized maximum likelihood estimation, which is repeated multiple times.
[0010] Preferably, in the process of establishing the LASSO-Cox regression model, the optimal regularization parameter λ is determined by the minimum value of the partial likelihood deviation, and then the optimal number of marker genes and regression coefficient are determined by the λ value.
[0011] The present invention also provides a prognostic model related to airspace dissemination of early-stage lung adenocarcinoma. The prognostic model is evaluated by risk scoring, and the risk score is calculated as follows: Risk score = (0.104059 × ID1 expression level) + (0.147979 × SLC16A1 expression level) + (0.151562 × FBXO17 expression level) + (-0.202947 × CRYL1 expression level) + (-0.585299 × NONO expression level) + (-0.314863 × MTIF3 expression level) + (0.415686 × SH3TC2 expression level).
[0012] The present invention also provides a computer program product, comprising a computer program, which, when executed by a processor, uses the prognostic model obtained by the above-mentioned construction method or uses the above-mentioned prognostic model for evaluation.
[0013] Beneficial effects: Based on early-stage lung cancer sequencing data, the key molecular signature of airway dissemination, a unique metastatic pattern in lung cancer, was identified, and a prognostic model accurately predicting patient survival was constructed based on this signature. Validated across multiple data sets, the results demonstrated that the airway dissemination-related risk model achieved the best prognostic stratification capability. Compared to clinical data, the seven-gene model had a higher consistency index (C-index) with survival. Furthermore, the gene combination exhibited significant superiority, achieving the best results compared to single genes and other gene combinations. This demonstrates that not only a single gene, but also not any arbitrary gene combination, can achieve the effects of the present invention. Therefore, the airway dissemination-related gene combination prognostic model of the present invention has improved predictive accuracy and universality, enabling clinicians to provide more accurate guidance for treatment decisions for patients with lung adenocarcinoma. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] Figure 1 This is a flowchart for constructing a prognostic model related to the airspace dissemination characteristics of early-stage lung adenocarcinoma according to Example 1 of the present invention; Figure 2 This is an example of the present invention regarding air cavity dissemination; Figure 3 This is the training set and validation-level screening process for constructing the prognostic model of airway dissemination-related genes in the present invention; Figure 4 Schematic diagram of coefficients for constructing a prognostic model of genes related to airway dissemination according to the present invention; Figure 5 Schematic diagram of parameters for constructing a prognostic model of genes related to airway dissemination according to the present invention; Figure 6 Schematic diagram of univariate COX regression analysis of the seven-gene signature risk score associated with airway dissemination in the training set of the present invention and clinically relevant characteristics; Figure 7 Schematic diagram of the ROC curve for the risk score of the seven-gene signature associated with air cavity dissemination in the training set of the present invention; Figure 8 is a schematic diagram of the survival curve of the seven-gene signature risk score associated with airway dissemination in the training set of the present invention; Figure 9 Schematic diagram of the validation of the seven-gene signature risk score associated with airway dissemination in the validation set GSE13213 of the present invention; Figure 10 This is a schematic diagram of the validation of the seven-gene signature risk score associated with airway dissemination in the validation set GSE26939 of the present invention; Figure 11 Schematic diagram of the validation of the seven-gene signature risk score associated with airway dissemination in the validation set GSE30219 of the present invention; Figure 12 This is a schematic diagram of the validation of the seven-gene signature risk score associated with airway dissemination in the validation set GSE72094 of the present invention; in, Figure 4 、 Figure 5 The horizontal axis is the value calculated by the mathematical formula and has no unit; Figure 6 、 Figure 7 The horizontal axis is the scale and has no unit; Figure 8 、 Figure 9 、 Figure 10 、 Figure 11 、 Figure 12 The horizontal axis is "time (year)". DETAILED DESCRIPTION
[0015] In order to make the purposes, features, and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0016] like Figures 1 to 12 As shown in Figure 2, a prognostic model for airspace dissemination in early-stage lung adenocarcinoma is constructed with a flowchart as shown in Figure 2. Figure 1 As shown, the following steps are included: 1) Dataset A lung adenocarcinoma cohort established at Jiangsu Cancer Hospital was used for model training and development. RNA expression profile data and corresponding clinical data of lung adenocarcinoma patients were retrieved from the Cancer Genome Atlas (TCGA) database. RNA expression profile data and clinical information of the validation cohort were obtained from the Gene Expression Omnibus (GEO) database (GSE13213, GSE26939, GSE30219, and GSE72094). 2) Genes related to airway dissemination characteristics Based on pathological slides from lung adenocarcinoma patients at Jiangsu Cancer Hospital, we annotated the characteristics of airspace dissemination in lung adenocarcinoma and categorized patients as those with positive or negative airspace dissemination. Based on the airspace dissemination signatures, we performed differential analysis on RNA expression profile data to identify a gene set associated with airspace dissemination.
[0017] 3) Construction of a prognostic model for genes associated with airway dissemination Based on the gene set associated with airway dissemination characteristics and corresponding clinical data, a LASSO-Cox regression model was established using the R package "glmnet" to screen marker genes and obtain the corresponding regression coefficient for each marker gene to form a prognostic model. In the process of establishing the LASSO-Cox regression model, penalized maximum likelihood estimation was performed using the bootstrap method with 5000 iterations. The optimal regularization parameter λ was determined by minimizing the partial likelihood deviation, and this λ value was then used to determine the optimal number of marker genes and regression coefficient. The prognostic model was evaluated using a risk score, which was calculated as follows: Risk Score =
[0018] Where n represents the total amount of marker genes; represents the expression level of the i-th marker gene; represents the regression coefficient of the i-th marker gene.
[0019] LUAD patients were divided into low-risk and high-risk groups based on the median risk score, and overall survival (OS) between the two groups was compared using Kaplan-Meier analysis. The R packages "survival," "survminer," and "timeROC" were used to plot receiver operating characteristic (ROC) curves and calculate the area under the curve (AUC) at 1, 3, and 5 years. Clinicopathological characteristics (sex, age, stage) and the risk score were included in a multivariate Cox regression analysis to verify whether the risk score in the prognostic model could serve as an independent risk factor for predicting overall survival. LUAD cohorts from the TCGA and GEO databases (GSE13213, GSE26939, GSE30219, and GSE72094) were used for validation, and the risk score was calculated using the same method as above, dividing the cohort into two subgroups (low-risk group and high-risk group). In the LASSO-Cox regression model, as the lambda value increases, the regression coefficient of the gene gradually shrinks. According to the results of 5000 bootstrap sampling, when the partial likelihood deviation takes the minimum value, the optimal solution of the model is obtained. At this time, 7 genes are obtained for prognostic model construction, namely ID1, SLC16A1, FBXO17, CRYL1, NONO, MTIF3, and SH3TC2. The regression coefficient corresponding to each gene is used to calculate the risk score, as shown in the figure below. Figure 2 , as shown in Figure 3.
[0020] In this embodiment, the prognostic model is specifically: Risk score = (0.104059 × ID1 expression level) + (0.147979 × SLC16A1 expression level) + (0.151562 × FBXO17 expression level) + (-0.202947 × CRYL1 expression level) + (-0.585299 × NONO expression level) + (-0.314863 × MTIF3 expression level) + (0.415686 × SH3TC2 expression level). Based on the calculated median risk score, the patients were divided into a high-risk group and a low-risk group. The OS of the high-risk group was significantly lower than that of the low-risk group (P < 0.0001). According to the ROC curve, the AUCs of the 1-year, 3-year, and 5-year survival prognostic models were 0.743, 0.712, and 0.690, respectively, indicating that the model had a good predictive effect; the univariate COX results showed that the risk score could be used as the best prognostic factor in patients with early-stage lung adenocarcinoma.
[0021] In the validation sets of GSE13213, GSE26939, GSE30219 and GSE72094, the OS of the high-risk group was significantly lower than that of the risk group, successfully verifying the universality of the present invention.
[0022] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.
Claims
1. A method for constructing a prognostic model related to airspace dissemination in early-stage lung adenocarcinoma, characterized by: The following steps are involved: S1: Obtain RNA expression profile data and corresponding clinical data of patients with lung adenocarcinoma; S2: Label the airway dissemination characteristics of lung adenocarcinoma patients and divide them into airway dissemination-positive and airway dissemination-negative patients. Based on the labeled airway dissemination characteristics, perform differential analysis on RNA expression profile data to obtain a gene set related to airway dissemination characteristics. S3: Based on the gene set related to airway dissemination characteristics and the corresponding clinical data, a regression model is established to screen out marker genes and obtain the corresponding regression coefficient of each marker gene, thereby forming a prognostic model.
2. The method for constructing a prognostic model related to airspace dissemination of early-stage lung adenocarcinoma according to claim 1, characterized in that: The prognostic model is evaluated by risk scoring, which is calculated as follows: Risk Score = Where n represents the total amount of marker genes; represents the expression level of the i-th marker gene; represents the regression coefficient of the i-th marker gene.
3. The method for constructing a prognostic model related to airspace dissemination of early-stage lung adenocarcinoma according to claim 1, characterized in that: The basis for marking the airspace dissemination characteristics of patients with lung adenocarcinoma is the pathological sections of patients with lung adenocarcinoma.
4. The method for constructing a prognostic model related to airspace dissemination of early-stage lung adenocarcinoma according to claim 1, characterized in that: The regression model is a LASSO-Cox regression model.
5. The method for constructing a prognostic model related to airspace dissemination of early-stage lung adenocarcinoma according to claim 4, characterized in that: In the process of establishing the LASSO-Cox regression model, the bootstrap method was used to perform penalized maximum likelihood estimation, which was repeated several times.
6. The method for constructing a prognostic model related to airspace dissemination of early-stage lung adenocarcinoma according to claim 4, characterized in that: In the process of establishing the LASSO-Cox regression model, the optimal regularization parameter λ is determined by the minimum value of the partial likelihood deviation, and then the optimal number of marker genes and regression coefficient are determined by this λ value.
7. A prognostic model for airspace dissemination in early-stage lung adenocarcinoma, wherein the prognostic model is evaluated by a risk score, and the risk score is calculated as follows: Risk score = (0.104059 × ID1 expression level) + (0.147979 × SLC16A1 expression level) + (0.151562 × FBXO17 expression level) + (-0.202947 × CRYL1 expression level) + (-0.585299 × NONO expression level) + (-0.314863 × MTIF3 expression level) + (0.415686 × SH3TC2 expression level).
8. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the prognostic model obtained by the construction method according to any one of claims 1 to 6 or the prognostic model according to claim 7 is used for evaluation.