Abstract: Lung adenocarcinoma(LUAD) is a disease with high morbidity and mortality rates, underscoring the need for enhanced accuracy and personalization in prognostic diagnostics. The integration of bioinformatics with positron emission tomography/computed tomography(PET/CT) presents substantial potential for the development of prognostic models and the identification of targeted therapeutics. In this study, we employed the Lasso-Cox algorithm to analyze LUAD transcriptomic data from the TCGA and GEO databases, successfully identifying prognosis-associated genes and constructing a high-performance prognostic prediction model. This model demonstrated robust stratification of high- and low-risk patients, achieving AUC values of 0.821, 0.693, and 0.701 for 1-, 3-, and 5-year survival rates, respectively. Subsequent virtual screening of the ChEMBL database using the model identified BAY-588 as a promising candidate compound. Computational simulations of ¹⁸F radiolabeling revealed its potential utility in PET imaging, providing scientific evidence for the development of theranostic radiopharmaceuticals targeting LUAD.