Identification of a New Eight-Long Noncoding RNA Molecular Signature for Breast Cancer Survival Prediction.

Abstract

Because of the phenotypic and molecular diversity, it is still difficult to predict breast cancer prognosis. This study aimed to develop and validate a multi-lncRNA (long noncoding RNA) signature to improve the survival prediction for breast cancer. Three hundred twenty-seven breast cancer patients from GSE20685 were used as a training set. GSE88770 including 117 patients and The Cancer Genome Atlas datasets including 1077 patients were used as 2 validation sets. Kaplan-Meier curve, the LASSO (least absolute shrinkage and selection operator) method, univariate and multivariate Cox analyses were applied to build a molecular model for predicting survival. Function analysis of this lncRNA signature was investigated. A novel eight-lncRNA molecular signature was first identified from multiple datasets. This signature classified patients into the high-risk and low-risk groups. Breast cancer in the high-risk group showed significantly worse survival than that in the low-risk group. Further multivariate Cox analysis revealed that this molecular signature was an independent prognostic factor for breast cancer in the training and validation sets. Furthermore, stratification analyses showed that this molecular signature was also used to classify patients into the low- and high-risk groups in patients with low or high T-stage, patients with or without lymph node metastasis, older or younger, estrogen receptor-positive or -negative, and progesterone receptor-positive or -negative patients. Our eight-lncRNA signature was a powerful tool in predicting prognosis in Luminal B breast cancer based on molecular subtype. This lncRNA signature involved in cell adhesion, apoptosis, cell differentiation, and immune regulation. Our study provided a reliable eight-lncRNA molecular signature for survival prediction of breast cancer, and this signature can stratify patients into the high- and low-risk groups. This molecular signature may help the selection of the suitable treatment strategies.