21O - A novel methylation signature that reflects intratumoral lymphocyte infiltration in breast cancer and predicts for response to anthracycline treatment
|Date||08 May 2014|
|Session||Best abstracts session|
|Topics|| Breast Cancer
|Citation||Annals of Oncology (2014) 25 (suppl_1): i8-i16. 10.1093/annonc/mdu066|
J. Jeschke1, M. Defrance1, C. Desmedt2, M. Bizet1, S. Dedeurwaerder1, E. Calonne1, C. Sotiriou3, F. Fuks1
Biomarkers that predict response or resistance to anthracycline treatment remain an unmet medical need in breast cancer. In this study we utilized ‘The neoadjuvant Trial of Principle’ (TOP) cohort, in which 149 patients with estrogen receptor (ER)-negative tumors were treated with anthracycline (epirubicin) monotherapy, to define methylation markers that reliably identify patients who do not benefit from anthracycline treatment.
DNA methylation profiles were obtained from 8 epithelial breast cancer cell lines and 4 normal lymphocyte cell lines using Infinium HumanMethylation27K arrays. DNA methylation profiles for 58 patients from the TOP cohort were assessed in pre-chemotherapy biopsies using Infinium HumanMethylation450K arrays. The primary endpoint of the study was pathologic complete response (pCR). A pCR was obtained in 7 out of 58 patients. Predictive methylation markers were identified in two steps: 1) Calculating differences in DNA methylation profiles between epithelial cancer cells and lymphocytes to identify differential methylated immune cell-related genes. 2) Testing the differential methylated immune cell-related genes for prediction of response in the TOP cohort. Focus was given to immune cell-related genes since tumor-infiltrating lymphocytes (TILs) have been previously identified as candidate biomarker for benefit to anthracycline-based chemotherapy.
We first developed a methylation signature including 10 immune cell-related genes (involved in T lymphocyte regulation), which were differentially methylated between normal lymphocytes and breast epithelial cells as a proxy of a quantitative measurement of TILs and antitumor immune response. We then showed that this signature was significantly associated with pCR with high negative predictive value (NPV, 0.78; 95% CI, 0.65 to 0.89) and a high positive predictive value (PPV, 0.86; 95% CI, 0.42 to 0.98).
If further validated, this methylation signature may be considered as a useful clinical tool to identify patients who will not benefit from anthracycline treatment and can be spared the burden of this therapy.
All authors have declared no conflicts of interest.