| HER-2 overexpression as a predictor of outcome in breast cancer patients
The HER-2 proto-oncogene is overexpressed/amplified in 20-30% of breast cancers that show a more aggressive behaviour. In patients with node positive breast cancer the HER overexpression is associated with a shorter RFS and OS and is an independent predictor of prognosis on multivariate analysis in some studies. In node negative patients conflicting results have been obtained with some studies indicating HER-2 as a prognostic factor and others not showing such correlation It
has been postulated that overexpression of HER-2 may modulate the clinical
sensitivity of tumors to chemotherapeutic or endocrine treatments. However,
clinical studies have yielded conflicting results. To address this issue,
we retrospectively evaluated HER2 expression and its interaction with
treatment in 266 of 348 patients with stage I/II breast cancer randomised
to receive adjuvant CMF (cyclophosphamide, methotrexate, 5-fluorouracil)
for 6 cycles or weekly epirubicin for 16 weeks. HER2 expression was
determined in a single reference laboratory by immunohistochemistry using
the monoclonal antibody CB11. Subsequently the same slides were reanalyzed
using the HerceptestÒ
score system .
We found that HER2 overexpression is a prognostic marker for poorer overall survival (OS) but not for relapse-free survival (RFS), on univariate analysis only. A Cox regression model did not confirm the prognostic role of HER2 for OS. These results are in agreement with the data of other studies that reported prognostic significance of HER2 expression on univariate analysis only. At 8 years, we did not observe any statistically significant difference in RFS and OS between the treatment arms in patients with low and high HER2 expression. However, when we analysed the outcomes of patients in the CMF arm we found that patients with HER2 overexpression had a worse OS than did patients without HER2 overexpression, whereas this difference was not significant in the epirubicin arm. These results may suggest that an antracycline-based therapy is more effective in patients whose disease overexpresses HER2. Our study, as well as other studies evaluating the predictive role of HER2, has several potential limitations: first, it is a retrospective study with incomplete evaluation of all patients included in the original adjuvant trial; second, it was not powered for comparison of HER2 overexpression in subgroups based on treatment regimen; third the relative low prevalence of HER2 overexpression further reduces the statistical power of a single study. Predictive strength of HER2 overexpression should be more appropriately evaluated in the scenario of randomised trials.
Reference:
Mariantonietta Colozza Medical Oncology Division, Azienda Ospedaliera, Sant'Andrea delle Fratte, Perugia, Italy
|