Abstract
BACKGROUND: Lung adenocarcinoma (LUAD) with brain metastasis (BM) occurs frequently and has a poor prognosis. In this study, we aimed to assess the correlation between gene expression signatures and the development of BM after surgical resection of LUAD. METHODS: We analyzed the immune-related gene expression profiles of 72 LUADs with and without BM after surgery and verified them using NanoString method and immunohistochemistry (IHC). We matched the Tumor, Node, Metastasis (TNM) stage in the groups with and without BM to minimize the effect of TNM stage. Pathway enrichment studies were also performed. RESULTS: In the NanoString results, we identified 11 upregulated immune-related gene signature that correlated specifically with BM in the discovery and validation sets [area under the curve (AUC) =0.750 and 0.787, respectively]. The discovery set achieved 94% sensitivity and 62% specificity and the validation set displayed 100% sensitivity and 50% specificity. Eight out of the 11 genes were verified by IHC and had profiles similar to the gene expression profile results (AUC =0.844 for the discovery set and AUC =0.795 for the validation set). Subgroup analysis revealed that 11 immune-related gene signature enabled prediction of BM at all TNM stages. There were no differences in the 11 immune-related gene expression signatures between the primary LUAD samples and the matched brain samples. Pathway enrichment analysis revealed that the cytokine-cytokine receptor interaction pathway was closely correlated with BM. CONCLUSIONS: The 11 identified immune-related gene expression signatures may be potentially clinically useful predictors for BM and can provide patient-specific treatment options.