Genomic aberrations in young and elderly breast cancer patients

Background: Age at breast cancer diagnosis is a known prognostic factor. Previously, several groups including ours have shown that young age at diagnosis is associated with higher prevalence of basal-like tumors and aggressive tumor phenotypes. Yet the impact of age at diagnosis on the genomic landscape of breast cancer remains unclear. In this study, we examined the pattern of somatic mutations, chromosomal copy number variations (CNVs) and transcriptomic profiles in young and elderly breast cancer patients. Methods: Analyses were performed on The Cancer Genome Atlas (TCGA) dataset. Patients with metastatic disease at diagnosis, classified as normal-like by PAM50 or had missing clinical information were excluded. Young patients were defined as <= 45 years of age, while elderly patients were those >= 70 years of age at breast cancer diagnosis. The remaining patients were classified as intermediate. We evaluated the association between age at diagnosis and somatic mutations, CNV and gene expression in a logistic regression model adjusting for tumor size, nodal status, histology and breast cancer subtype. All analyses were corrected for multiple testing using the Benjamini-Hochberg approach. Results: In this study, 125, 486 and 169 patients were <= 45, 46-69 and >= 70 years of age, respectively. Older patients had more somatic mutations (n = 44 versus 35 versus 31; P = 0.0009) and more CNVs, especially in ductal tumors (P = 0.02). Eleven mutations were independently associated with age at diagnosis, of which only GATA3 was associated with young age (15.2 % versus 8.2 % versus 9 %; P = 0.003). Only two CNV events were independently associated with age, with more chr18p losses in older patients and more chr6q27 deletions in younger ones. Younger age at diagnosis was associated with higher expression of gene signatures related to proliferation, stem cell features and endocrine resistance. Conclusions: Age adds a layer of biological complexity beyond breast cancer molecular subtypes, classic pathological and clinical variables, worthy of further consideration in future drug development as we seek to refine therapeutic strategies in the era of personalized medicine.