A multi-stage computational pipeline for repurposing FDA-approved drugs: application to EGFR C797S–mutant NSCLC

Background/Objectives Non-small cell lung cancer (NSCLC) is a leading cause of cancer-related death globally. Resistance to third-generation epidermal growth factor receptor (EGFR) tyrosine kinase inhibitors, particularly due to the C797S mutation, poses a significant clinical challenge. This study utilized a comprehensive multiphase computational drug repurposing approach to discover FDA-approved medications that may be effective against the EGFR C797S mutation. Methods A library of 1,650 compounds from the ZINC15 database was subjected to shape-based screening, followed by hierarchical molecular docking using high-throughput, standard precision, and extra-precision methods. The top candidates were further analyzed using MM-GBSA binding free-energy calculations, covalent docking simulations, and three 300-ns molecular dynamics simulations to evaluate the binding stability and interaction persistence under dynamic conditions. Results Among the evaluated compounds, doripenem, norgestrel, oxymetholone, norethisterone, and ertapenem exhibited high docking scores and consistent interactions with crucial hinge-region residues, such as MET-793 and mutant SER-797, within the ATP-binding site. Through molecular dynamics analyses, ertapenem and oxymetholone were identified as the most stable complexes, exhibiting minimal root-mean-square deviation fluctuations and maintaining hydrogen-bond networks similar to that of the reference inhibitor osimertinib. Conclusion These findings suggest that ertapenem and oxymetholone are promising structurally unique scaffolds for targeting osimertinib-resistant EGFR C797S–driven NSCLC. While experimental validation is necessary, this study provides a computational framework for swift drug repurposing and lays a rational foundation for future biochemical evaluation and structure-guided optimization of next-generation EGFR inhibitors.