BACKGROUND: There is an unmet need for more efficient patient stratification for receiving trastuzumab in the metastatic breast cancer (mBC) setting, since only part of such patients benefit from the addition of this agent to chemotherapy. The aim of this study was to investigate the prognostic value of biomarkers including MYC and MET in mBC patients treated with trastuzumab-based regimens.
METHODS: mBC patients, locally tested as HER2-positive, treated with trastuzumab and chemotherapy between 1998 and 2010 were evaluated. Paraffin tumors (n = 229) were retrospectively centrally assessed by immunohistochemistry (IHC) for HER2, ER, PgR and Ki67; fluorescence in situ hybridization (FISH) for HER2, TOP2A and centromere (CEN) 17, MYC and CEN8, MET and CEN7; qPCR for MYC, MET copy number (CN); and, for PI3K activation (PIK3CA mutations; PTEN and phospho-mTOR protein expression). Increased CEN CN was assessed based on normal cut-offs. Time to progression (TTP) and survival were evaluated from the initiation of trastuzumab as first line treatment.
RESULTS: Among all tumors, 90 were HER2-negative upon central testing (ambiguous HER2) and the rest were true HER2-positive. Further, 156 patients presented with mBC upon relapse of pre-treated disease (R-mBC) and 65 were diagnosed at stage IV (de novo mBC). Concordance between FISH and qPCR on gene CN status was fair for MYC (Kappa = 0.458) and absent for MET. The presence of MYC CN gain with qPCR and the absence of PI3K activation were infrequent events (7 and 8 % of evaluable tumors, respectively), while 41 % of tumors had increased CEN CN in one or more chromosomes, indicative of chromosomal instability. The most consistent finding in the entire cohort and in the above patient subgroups with respect to outcome was the unfavourable effect of MYC CN gain, which was retained upon multivariable analysis (e.g., survival in the entire cohort, HR 6.02; 95 % CI 2.67-13.6; p < 0.001). Further unfavourable prognosticators were increased CEN CN in one chromosome in R-mBC but not in de novo mBC (multivariable interaction p = 0.048), PI3K activation in R-mBC (multivariable p = 0.004) and increased Ki67 for patient TTP.
CONCLUSIONS: MYC gene copies, centromere status and PI3K activation may adversely impact trastuzumab treated mBC patient outcome and seem worthy validating in larger series.