The extended concurrent genes signature for disease-free survival in breast cancer

• Published in: Journal of the Formosan Medical Association Vol. 121; no. 10; pp. 1945 - 1955
• Main Authors: Huang, Ching-Shui, Tsai, Ming-Lin, Lu, Tzu-Pin, Tu, Chao-Chiang, Liu, Chih-Yi, Huang, Chi-Jung, Ho, Yuan-Soon, Tu, Shih-Hsin, Chuang, Eric Y., Tseng, Ling-Ming, Huang, Chi-Cheng
• Format: Journal Article
• Language: English
• Published: Singapore Elsevier B.V 01-10-2022; Elsevier
• Subjects: Breast cancer; Consensus; Extended concurrent genes signature; Leading edge analysis; Partial least square regression; Breast cancer; Leading edge analysis; Consensus; Extended concurrent genes signature; Partial least square regression
• ISSN: 0929-6646; 1876-0821
• DOI: 10.1016/j.jfma.2022.01.022

Previously we had identified concurrent genes, 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 to discovery biomarkers not identifiable by conventional GE only data, for which phenotype-correlation or gene variability is the criteria of gene selection.
Thirty-one comparative genomic hybridization (CGH) and 83 GE microarrays were performed, with 29 breast cancers assayed from both platforms. Potential targets were revealed by Genomic Identification of Significant Targets in Cancer (GISTIC) from CGH arrays. Concurrent genes and genes with significant GISTIC scores were used to derive the extended concurrent genes signature, which was consensus from leading edge analysis across all studies and a supervised partial least square (PLS) regression predictive model of disease-free survival was constructed.
There were 1584 concurrent genes from 29 samples with both CGH and GE microarrays. Enriched concurrent genes sets for disease-free survival were identified independently from 83 GE arrays and another one with Han Chinese origin as well as three studies of Western origin. For five studies with disease-free survival follow up, prognostic discrepancy was observed between predicted high-risk and low-risk group patients.
We concluded that through parallel analyses of CGH and GE microarrays, the proposed extended concurrent gene expression signature can identify biomarkers with prognostic values.