Abstract
Aims
This study aimed to develop accurate prognostic assessment tools for patients with lung adenocarcinoma (LUAD) using Summary-data-based Mendelian Randomization Analysis and statistical learning-based prognostic modeling methods.Methods
We conducted SMR analysis to explore genetic associations between genes and LUAD, and analyzed differentially expressed genes (DEGs). The shared genes were subjected to the univariate Cox regression analysis and survival analysis to determine prognosis-related genes to construct a prognostic model with the multivariate Cox regression analysis. Individual gene contributions to the model were clarified through SHAP analysis, and predictive performance was assessed using the receiver operating characteristic curve, the C-index, and calibration plots.Results
59 shared genes were identified between the SMR and DEGs analyses. The univariate Cox regression analysis and survival analysis identified two prognosis-related genes (PKP2 and SLC7A11), which were used to develop a prognostic model. The Kaplan-Meier plot and univariate and multivariate Cox regression models illustrated that patients in the high-risk group had a poorer clinical outcome. The low-risk group showed enrichment of several immune-related pathways and showed higher infiltration levels of 15 immune cell types.Conclusion
A two-gene prognostic model was developed to predict outcomes in patients with LUAD, which was also highly associated with immune status.