A Nomogram Combining Two Novel Biomarkers for Predicting Lung Adenocarcinoma in Ground-Glass Nodule Patients

Abstract

Objective: Combination of CT imaging and RNA sequencing techniques was used to explore the potential biomarkers specific to lung adenocarcinoma within pulmonary ground-glass nodules. Method: The imaging and pathological data of patients with pulmonary ground-glass nodules who underwent chest CT scanning were confirmed through surgical procedures. Based on the pathological results, the patients were categorized into a benign nodule group and a malignant nodule group. Subsequently, RNA sequencing was conducted to analyze gene expression information in the pulmonary ground-glass nodules of these 16 patients. Results: CT signs demonstrated statistical significance in both benign and malignant nodules. A total of 2080 upregulated genes and 1240 downregulated genes were identified through RNA sequencing in malignant nodules compared to benign nodules. CST1 exhibited increased expression among the upregulated genes in lung adenocarcinoma tissues compared to lung tissues. Among the downregulated genes, only GIMAP1-GIMAP5 showed decreased expression in lung adenocarcinoma tissues. Finally, we validated the clinical significance of CST1 and GIMAP1-GIMAP5 in patients with lung adenocarcinoma, particularly highlighting a strong correlation between GIMAP1-GIMAP5 expression levels and prognosis for patients. A visual nomogram predictive model for pulmonary ground-glass nodules was constructed (area under the receiver operating characteristic curve (AUC) > 0.8). Conclusion: We constructed a nomogram combining CST1 and GIMAP1-GIMAP5 expression for predicting lung adenocarcinoma in ground-glass nodules in the context of COVID-19. This nomogram addresses the unique diagnostic challenges posed by COVID-19, where overlapping pulmonary imaging findings between viral pneumonia and early lung cancer necessitate robust molecular-aided discrimination.

Keywords: RNA sequencing technology; lung adenocarcinoma; nomogram; pulmonary ground-glass nodules.

Figure 1

Study flowchart.

Figure 2

Typical imaging-pathological data presentation. (a) Benign nodule, (b) AIS, (c) MIA, and (d) IPA.

Figure 3

The differential expression of genes between benign and malignant nodules. (a) Principal component analysis. (b) Volcano plot displaying the fold change and q value of coding genes between benign and malignant nodules. The x-axis represents the fold change in gene expression, which is considered biologically significant only when the threshold of change is set at Log2FC > 1 or < −1. The y-axis represents statistical significance, with a threshold set at a q value of 0.05 (−Log10q value > 1.3). Upregulated genes, indicated by red dots, have a fold change greater than 2 and p value less than 0.05. Downregulated genes, shown as green dots, have a fold change less than 0.5 and p value less than 0.05; genes with nonsignificant differences are represented by gray dots. (c) Clustering analysis of differentially expressed genes: The x-axis represents sample names between groups, while the y-axis represents differentially expressed genes. In this figure, high expression values of differentially expressed genes in group samples are denoted by red dots, whereas low expression values are denoted by green dots.

Figure 4

Functional and signaling pathway analysis of significantly differentially expressed genes. (a) Functional histogram of significantly upregulated differentially expressed genes. (b) Functional histogram of significantly downregulated differentially expressed genes. In the figure, the vertical axis represents the functional names of differentially expressed genes, and the horizontal axis represents the negative logarithm of p value (−LgP). A higher value indicates a smaller p value, indicating a greater significance level for the function of differentially expressed genes (only displaying the top 10 items in descending order based on −LgP values). (c) Bubble chart illustrating significant pathways enriched with upregulated differentially expressed genes. (d) Bubble chart illustrating significant pathways enriched with downregulated differentially expressed genes. In the figure, the vertical axis represents the pathway names for differentially expressed genes, and the horizontal axis represents the negative logarithm of p value (−LgP). A higher value indicates a smaller p value, indicating a greater significance level for that particular pathway. The size of each bubble corresponds to the number of differentially expressed genes within that pathway. (e) Interaction network diagram depicting significant pathways. In this diagram, circles represent individual pathways while straight lines indicate interactions between them. Red circles represent pathways associated with upregulated gene expression, blue circles represent pathways associated with downregulated gene expression, and yellow circles represent pathways exhibiting both up and downregulation properties.

Figure 5

Expression patterns of CST1 and GIMAP1-GIMAP5 in lung adenocarcinoma and their prognostic implications. (a) GSE118370 and GSE140343 datasets were selected for including gene expression profiles of lung adenocarcinoma tissues and normal lung tissues to validate differentially expressed genes identified in our study. The Venn diagram illustrates the intersection of datasets GSE118370, GSE140343, and the Top 10 upregulated or downregulated genes. (b) Within the GSE118370 dataset, expression of CST1 and GIMAP1-GIMAP5 in lung cancer tissues. (c) Expression of the CST1 gene and GIMAP1-GIMAP5 gene in lung adenocarcinoma tissues. (d) Correlation between the expression level of CST1, GIMAP1-GIMAP5 and the prognosis of patients with lung adenocarcinoma.

Figure 6

Lung cancer risk nomogram predictive chart. (a) Predictive nomogram model incorporating five variables (gender, nodule size, CST1, and GIMAP1-GIMAP5) for estimating the probability of lung cancer using a binary logistic regression framework. (b) Receiver operating characteristic (ROC) curve of the binary logistic regression model developed with gender, nodule size, CST1, and GIMAP1-GIMAP5 to predict lung cancer probability. (c) Precision–recall (PR) curve of the prediction model integrating gender, nodule size, CST1, and GIMAP1-GIMAP5 to assess the likelihood of lung cancer.