Abstract
BACKGROUND: Comprehensive genomic analysis and optimal treatment for non-small cell lung cancer (NSCLC) patients with epidermal growth factor receptor (EGFR) G719X + S768I co-mutations remain limited. This study aimed to elucidate the genetic landscape and the clinical effectiveness of EGFR tyrosine kinase inhibitors (EGFR-TKIs) in this subset. METHODS: A total of 645 EGFR-mutant NSCLC patients were retrospectively screened, with 142 patients harbored EGFR G719X + S768I co-mutations. Among these patients, next-generation sequencing was performed in 126 patients, and the efficacy of first-line EGFR-TKIs was evaluated in 96 patients with stage IV disease. Impacts of variant allele frequency (VAF) and concurrent TP53 mutations were also analyzed. RESULTS: Among G719X variants, G719C was most prevalent (69.8%), followed by G719A (19.0%) and G719S (8.7%). The most common co-existing mutation was TP53 (38.0%), followed by ALK and PIK3CA (6.3% each). For first-line EGFR-TKIs, the overall objective response rate (ORR) reached 68.8%, with a median progression-free survival (mPFS) of 21.4 months (95% CI: 18.2–24.6). Afatinib showed a significantly better response compared to both first- and third-generation EGFR-TKIs (1st vs. 2nd vs. 3rd generations, ORR: 35.7% vs. 76.8% vs. 61.5%, P = 0.01; mPFS: 17.2 vs. 23.4 vs. 17.4 months, P = 0.008). VAF and TP53 mutation status did not affect outcomes (P > 0.05), but patients with metastases in the brain and liver experienced notably shorter mPFS. Brain metastases patients had an ORR of 70.5% without additional benefit from third-generation TKIs. CONCLUSIONS: This study delineates the genomic profile of EGFR G719X + S768I co-mutated NSCLC and highlights the superior effectiveness of second-generation EGFR-TKIs, particularly afatinib. These findings provided valuable insights to guide clinical decision-making and facilitate the development of tailored therapeutic strategies for this subset. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12885-026-15887-6.