Background: Histological grade (HG) is one the main prognostic factors used in clinical practice, along with nodal status and tumour size, for the classification of patients into different prognostic groups that support the clinical decision making process and to tailor adjuvant treatment according to specific risk of death or relapse. The aim of this study was to examine the effect of Grade in prognosis in an unselected series of breast cancers, and to examine copy number alteration (CNA) specifically in Grade 3 (G3) breast cancer and its relationship to nodal status (N-ve/N+ve) and clinical outcome, to seek genetically determined markers of prognosis for treatment tailoring in this poor prognosis group.Method: We established a clinical database of 1489 operable Breast Cancer patients attending the Oncology Service at Singleton Hospital, Wales, UK between 1996 and 2002. HG was known for 1339 of these cases, and outcome in 1213, with a median of 5.32 years (range 0.09 - 10.14 years) follow-up. From the 466 G3 cases of primary breast cancer, we analysed 81 cases using 1Mb BAC array CGH for genomic copy number alterations. Spatial normalisation, circular binary segmentation and the CGHcall algorithm was used to generate CGH profiles. Unsupervised hierarchical clustering, supervised, correlation and survival analyses were carried out using packages and tests within the R statistical platform.Results: Patients were characterised according to grade, ER and nodal status. Kaplan Meier analysis for overall survival (OS) and disease specific survival (DFS) confirmed the poor prognosis of G3 (Cox proportional hazards regression, Wald-test p-value << 0.05). The majority of G1 (98%) and G2 (93%) cases were ER +ve, whereas 52.5% G3 cases were ER-ve. ER status was known for 316 of the 466 G3 cases, and ER negativity was significantly associated with a decreased OS (p-value 0.02), but not DFS (p-value: 0.13). Similarly N+vity also showed significant correlation with both OS and DFS (p-values < 0.05). BAC aCGH analysis showed significant copy number gains on chromosomes 1,4,5,7,11 and 21 in N+ve and losses on chromosomes 1,6 and 10 in N-ve G3. One region of chromosome 1 (1p36 – 1p35) in particular was lost in N0 but gained in N+ve, and this was significantly associated with ER-ve/N+ve cases (FDR p-value < 0.05). BAC aCGH profiles did not significantly associate with ER status in G3 disease. Two clusters, delineated on unsupervised hierarchical cluster analysis, correlated with OS. Neither correlated with nodal or ER status, indicating that further analysis may discover genes of prognostic potential within these profiles.Conclusion HG was significantly associated with clinical outcome and ER status. Genomic copy number analysis (CNA) in G3 breast cancer revealed significant associations between DNA alterations and nodal status and OS. The study is currently being extended to investigate the relationship between specific CNA and clinical outcome in G3 disease.

Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 5167.

2009