BACKGROUND: The discovery of driver genes such as , , and has dramatically shifted treatment patterns in patients harboring these oncogenes. However, dissemination into the central nervous system (CNS) is a severe complication. In addition, the particular anatomical structure of the CNS has made it difficult to obtain tissue specimens from brain metastases (BM) to generate a gene map, as such, potential predictive markers for survival in patients with non-small cell lung cancer (NSCLC) and BM (NSCLC-BM) remain unclear. METHODS: Data from 28 patients diagnosed with NSCLC-BM between June 2019 and May 2021 at Guangdong Sanjiu Brain Hospital (Guangzhou, China), were reviewed. Targeted next-generation sequencing (NGS) of a 168 cancer-related gene panel was available for surgically resected brain tissues from all patients. In addition, molecular characteristics and overall survival (OS) were analyzed to determine potential predictive markers. RESULTS: Among patients with NSCLC-BM, NGS revealed that was the most frequent mutation (61 %), with a detection rate of 39 %, closely by amplification. Additionally, , , , and were frequently observed (18 %). The median OS was significantly shorter in the mutation group than in the wildtype group (14 versus undefined months,  = 0.014). Similar results were also found in the genetic alteration of amplification, suggesting that amplification was associated with worse OS (14 vs. 24 months,  = 0.039). Interestingly, NGS revealed that gene alternations such as , amplification, and , tended to coexist and such a co-alteration panel indicated worse clinical outcomes (median OS, 5 months). In addition, the detection rate of negative survival genes, including or amplification, was much higher in tumor tissues than in plasma samples, indicating the limited predictive value of matched PLA samples. CONCLUSIONS: Gene signatures, such as or amplification, were associated with worse survival in patients diagnosed with NSCLC-BM. These valuable findings may shed light on new strategies for the prognostic assessment of specific patient groups.