Archives

  • 2019-07
  • 2019-08
  • 2019-09
  • 2019-10
  • 2019-11
  • 2020-03
  • 2020-07
  • 2020-08
  • 2021-03
  • br Abbreviations aCGH array comparative genomic hybridizatio

    2019-09-23


    Abbreviations: aCGH, array comparative genomic hybridization; AUC(t), time-dependent area under the ROC curve function; C-index, concordance index; CI, confidence interval; CIN25/70, chromosomal instability-related gene sets; CINSARC, complexity index in sarcomas; CNA, copy number alteration; DFS, disease-free survival; DSB, double-strand break; DSS, disease-specific survival; ER, oestrogen receptor; G2I, genomic instability index; GGI, genomic grade index; HER2, human epidermal growth factor receptor 2; HR, hazard ratio; mRNA-seq, mRNA sequencing; OS, overall survival; PR, progesterone receptor; qRT-PCR, quantitative real-time reverse-transcriptase polymerase chain reaction; RFS, recurrence-free survival; SSC, stemline scatter index; t-SNE, t-distributed stochastic neighbour embedding; TCGA, The Cancer Genome Atlas
    Corresponding author at: Department of Oncology, Institute of Clinical Sciences, Sahlgrenska Cancer Center, Sahlgrenska Academy at University of Gothenburg, Box 425, SE-405 30 Gothenburg, Sweden.
    E-mail address: [email protected] (J. Biermann).
    1 Equal contributors.
    breast cancer patients with poor clinical outcome.
    Aneuploidy is one of the most common aberrations in cancer and potentially caused by genomic instability [16]. Therefore, identifying genes associated with genomic instability has the potential to optimise outcome prediction and minimise the risk of toxic side effects from chemotherapy by avoiding over-treatment. Here, we validated genomic instability as a prognostic marker defined on a genomic and tran-scriptomic level. A total of 136 breast carcinomas were stratified based on genome stability into genomically stable and unstable tumours using the genome instability index (G2I) [17]. Matching gene SC 236 profiles were analysed for differentially expressed transcripts using lo-gistic regression. In random forest models, the 335 differentially ex-pressed transcripts were further narrowed down to a 17-marker panel comprised of genes with the highest variable importance in distin-guishing genomically stable from unstable tumours. The 17-marker panel was externally validated and showed increased predictive power compared to the tubulins Oncotype Dx-based signature and the 12-gene genomic instability signature.
    2. Patients and methods
    2.1. Patients and clinicopathological data
    In total, 136 primary invasive breast carcinomas were selected from previously analysed patient cohorts mainly consisting of luminal B tu-mours from patients diagnosed in Western Sweden between 1991 and 1999 (Table 1) [18–20]. All patients were metastasis-free at the time of diagnosis. Primary treatment of the breast carcinomas was surgery followed by adjuvant therapy (radiotherapy, chemotherapy and/or endocrine therapy) when appropriate. Fresh-frozen tumour samples were stored in the tumour bank at the Sahlgrenska University Hospital Oncology Lab (Gothenburg, Sweden). Clinicopathological information were obtained from Regional Cancer Centre West (Gothenburg, Sweden). The dataset was stratified into the molecular breast cancer subtypes (normal-like, basal-like, luminal A, luminal B, luminal B HER2-amplified, and HER2/ER−) and genomic grade index (GGI; low, high) as described elsewhere [21–23]. Survival time was defined as the time from initial diagnosis to death from any cause (overall survival; OS) and the time from initial diagnosis to breast cancer-related death (disease-specific survival; DSS). The study was approved by the Re-gional Ethical Review Board in Gothenburg, Sweden. r> External validation was performed in three independent cohorts. The publicly available breast cancer datasets GSE1456 (n = 159) [24] and GSE4922 (n = 249) [25] were profiled with the Affymetrix Human Genome U133 Set and provided information on OS and disease-free survival (DFS; time from initial diagnosis to first relapse or breast cancer-related death), respectively. The TCGA dataset (The Cancer Genome Atlas Breast Invasive Carcinoma dataset [26]) consisted of mRNA sequencing (mRNA-seq) data for 899 primary breast tumours and contained survival time given as OS.