Abstract
Our recent work had identified concurrent gene signatures which highlighted the interplay between copy number variation (CNV) and differential gene expression (GE) for Han Chinese breast cancers. The merit of the approach is the discovery of biomarkers not readily identified by conventional GE only data, for which phenotype-correlation or gene variability is the criteria of gene selection.
A total of 31 array CGH and 83 GE arrays (29 samples with data from both platforms) were performed. Potential targets of cancer therapy were revealed by Genomic Identification of Significant Targets in Cancer (GISTIC) from 31 array CGH. We used these concurrent genes as well as genes with significant GISTIC scores to derive signatures associated with clinical ER, HER2 status, and disease-free survival.
The most common gain regions were 1q21.2 and 17q12-q21.1 while the most common lost region was 22q11.23. In subgroup analysis, recurrent gains for ER positive cancers were 8p23.1, and 8p11.23-11.21. For HER2 overexpressing tumors, recurrent gain was 17q12-21.1 and recurrent losses were 8p11.22, 8p12, and 8p23.1. We identified 1,584 concurrent genes from 29 samples assayed for both array CGH and GE microarrays, and enriched concurrent gene sets for ER, HER2 and disease-free survival were identified independently from our 83 GE arrays (GSE48391) and two series with Han Chinese origin (GSE5460 and GSE20685) as well as three studies of Western countries (GSE7390, GSE2034 and GSE3494). Signature genes were consensus genes from leading edge analysis across all studies, and we used supervised partial least square (PLS) regression to derived predictive model of clinical ER, HER2 and disease-free survival with each signature gene further weighted by the coefficient of concurrence as an additive tuning parameter. The predictive accuracy was 92.3% for ER and 84.6% for HER2 in 208 Han Chinese breast cancers and was 87.3% for ER in 735 breast cancers in Western countries. For five studies with disease-free survival data (no survival data for GSE5460), prognostic discrepancy was observed in four studies except GSE3493 between high-risk and low-risk patients predicted with the concurrent signature.
We conclude that parallel analysis of CGH and microarray data, in conjunction with known gene expression patterns, can be used to identify biomarkers with prognostic values in breast cancer.
Citation Format: Chi-Cheng Huang, Shih-Hsin Tu, Ching-Shui Huang, Heng-Hui Lien, Liang-Chuan Lai, Eric Y. Chuang. Extended concurrent gene signatures of ER, HER2 and disease-free survival in breast cancers. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 1400. doi:10.1158/1538-7445.AM2014-1400