Prediction of prognostic signatures in triple-negative breast cancer based on the differential expression analysis via NanoString nCounter immune panel

ABSTRACT: Triple Negative Breast Cancer (TNBC) has been considered an aggressive and complex subtype of breast cancer. Current biomarkers used in breast cancer treatment are highly dependent on targeting ER, PR, or HER2 in clinical practice, which results in ttreatment failure and disease recurrence continue to be clinically challenging. In this regards, there is still a crucial need for improvement of TNBC treatment by discovery of effective biomarkers that can be easily translated to the clinics and possible targets as prognosis and novel therapies. This study reports an approach for biomarker discovery, which predicts tumour relapse and pathologic complete response (pCR) in TNBC on the basis of mRNA expression quantified using NanoString nCounter Immunology Panel. We identify nine and 13 differentially expressed genes (DEGs), respectively, that are strongly associated with pCR and relapse from 579 immune genes over a small number of samples (n=55) using edgeR. To overcome a small sample size limitation, prediction models based on the random forest are constructed on the DEGs as selected features extracted from the DEG analysis. Comprehensive analysis indicated that our prediction models outperform those constructed on features extracted from the existing feature selection model such as, Elastic Net in terms of accuracy. The prediction models are assessed by randomization test to validate the robustness (empirical P for the model of pCR= 0.022 and empirical P for the model of relapse= 0.025). Furthermore, three DEGs (IL17B, EDNRB, and TGFBI) in the model of relapse show prognostic significance to predict cancer patient survival through Cox proportional hazards regression model based survival analysis.

ORGANISM(S): Homo sapiens

PROVIDER: GSE143222 | GEO | 2020/01/08

REPOSITORIES: GEO