Abstract P4-04-13: Could genes associated with response or resistance to radiotherapy be identified in breast cancer patients?

Abstract

Abstract Background. An absolute 10-year reduction of locoregional/distant relapse and 15-year reduction of mortality from breast cancer (BC) after breast conserving surgery by radiotherapy (RT) are about 16% and 3.8%, respectively (EBCTCG, Lancet 2011;378:1707–16). Similarly, after radical mastectomy absolute 5-year reduction of local recurrence by RT is 17% and 15-year BC mortality was reduced by 5.4% (EBCTCG, Lancet 2005,366:2087–2106). Predictive factors of response to RT are less frequently investigated compared to predictive factors for systemic therapy. The ideal model for researching the predictive value of markers is a neo-adjuvant setting. We have analyzed genes associated with different response to RT using neo-adjuvant model in locally-advanced (LA) BC patients (pts). Patients and methods. This retrospective analysis included LA non-inflammatory BC pts treated with preoperative (P)RT alone with a total dose of 45 Gy in 15 fractions every second day to the breast and regional lymph nodes. Radical mastectomy was performed 6 weeks after PRT to all pts and adjuvant systemic therapy was administered as per protocol. Response to PRT was classified as pathological complete response (pCR) if there is no evidence of tumor in the breast after mastectomy, partial response (PR) if regression of tumor was >30%, and no response if regression was <30%. We have analyzed global gene expression profiles by microarray analysis in a set of 14 samples of formalin-fixed paraffin-embedded (FFPE) invasive breast cancer tissues. Results. We have identified a group of 134 LABC pts treated with PRT between 1997 and 2000, which was the routine practice at that time according to the protocol for diagnosis and treatment of BC. Fourteen FFPE tumor samples, 7 from responders to PRT and 7 from non-responders to PRT, were collected for gene expression profile analysis. Unsupervised clustering analysis (Pearson correlation, un-centered metrics) over top 20% most variable genes, revealed clear separation of non-responders and responders. Unsupervised clustering analysis of paired pre- and post-PRT samples indicating small differences in expression profile in response to PRT within non-responders that were clustering together, unlike responders that were separated in the unsupervised analysis showing grater differences in their gene expression. We have found 98 genes significantly differentially expressed [Fals Discovery Rate (FDR) < 0.05] between responders and non-responders to PRT. Molecular functions significantly associated with this gene list were determined using Ingenuity Pathway Analysis and included regulation of cellular growth and proliferation, cell cycle, cell death and survival as top scoring function. Interestingly, canonical pathways for G2/M DNA damage checkpoint regulation and GADD45 genotoxic stress signaling were significantly enriched (FDR<0.05) among differentially expressed gene list between RT resistant and RT sensitive tumors. Conclusion: The strength of these results is that, although obtained on a small number of pts, these tumours have been exposed only to RT without systemic therapy administered either previously or concomitantly with PRT. Further confirmatory research of potentially RT sensitive/resistant genes is warranted. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-04-13.