Construction of a human epidermal growth factor receptor 2-related gene risk model for predicting breast cancer prognosis.

The present study aimed to construct a human epidermal growth factor receptor 2 (HER2)-related gene risk model to predict breast cancer prognosis. Gene expression and clinical follow-up data were extracted from The Cancer Genome Atlas database, while the GSE7390 dataset was obtained from the Gene Expression Omnibus database. Prognostic and clinical feature analyses were performed. In addition, differentially expressed genes (DEGs) between HER2-negative and -positive groups were screened, followed by enrichment analysis. Subsequently, a prognostic model was established, and prognosis was predicted using a nomogram. In addition, the association of risk score with immunity was analyzed, and single-cell analysis was performed. Next, key genes were identified by reverse transcription-quantitive PCR (RT-qPCR) analysis. The results revealed that HER2 was significantly associated with estrogen receptor status, progesterone receptor status, N stage, American Joint Committee on Cancer stage, mutation count and tumor mutation burden of breast cancer. AS601245, AP.24534 and roscovitine were the top three chemotherapeutic agents showing the highest sensitivity differences between the risk groups. A total of 251 DEGs between HER2-negative and -positive groups were screened, which were found to be significantly involved in the Kyoto Encyclopedia of Genes and Genomes pathway of estrogen signaling, PI3K-AKT signaling pathway and chemical carcinogenesis-receptor activation. Eight prognostic gene models were constructed, and it was found that patients in the high-risk group had significantly shorter survival times than those in the low-risk group. A nomogram, incorporating risk groups and clinicopathological features, demonstrated strong predictive ability and high accuracy. The RT-qPCR results indicated that the expression of electron transfer flavoprotein subunit α, rap guanine nucleotide exchange factor-like 1, keratin 7, cluster of differentiation 24, proline rich 15-like, arachidonate 15-lipoxygenase type B, ELOVL fatty acid elongase 2 and C-X-C motif chemokine ligand 9 was consistent with the results of bioinformatic analysis. In conclusion, the HER2-related risk model and nomogram developed in the present study demonstrated high accuracy in predicting patient survival.