6-lncRNA Assessment Model for Monitoring and Prognosis of HER2-Positive Breast Cancer: Based on Transcriptome Data

Abstract

Background: In view of the high malignancy and poor prognosis of human epidermal growth factor receptor 2 (HER2)-positive breast cancer, we analyzed the RNA expression profiles of HER2-positive breast cancer samples to identify the new prognostic biomarkers. Methods: The linear fitting method was used to identify the differentially expressed RNAs from the HER2-positive breast cancer RNA expression profiles in the Cancer Genome Atlas (TCGA). Then, a series of methods including univariate Cox, Kaplan-Meier, and random forests, were used to identify the core long non-coding RNAs (lncRNAs) with stable prognostic value for HER2-positive breast cancer. A clinical feature analysis was performed, and a competing endogenous RNA network was constructed to explore the role of these core lncRNAs in HER2-positive breast cancer. In addition, a functional analysis of differentially expressed messenger RNAs in HER-2 positive breast cancer also provided us with some enlightening insights. Results: The high expression of four core lncRNAs (AC010595.1, AC046168.1, AC069277.1, and AP000904.1) was associated with worse overall survival, while the low expression of LINC00528 and MIR762HG was associated with worse overall survival. The 6-lncRNA model has an especially good predictive power for overall survival (p < 0.0001) and 3-year survival (the area under the curve = 0.980) in HER2-positive breast cancer patients. Conclusion: This study provides a new efficient prognostic model and biomarkers of HER2-positive breast cancer. Meanwhile, it also provides a new perspective for elucidating the molecular mechanisms underlying HER2-positive breast cancer.