Comparison of different approaches for assessment of HER2 expression on protein and mRNA level: prediction of chemotherapy response in the neoadjuvant GeparTrio trial (NCT00544765)

Human epidermal growth factor receptor 2 (HER2) testing is an essential part of pathological assessment in breast cancer patients, as HER2 provides not only prognostic but also predictive information on response to targeted therapy. So far, HER2 test accuracy of immunohistochemistry/in situ-hybridization techniques is still under debate, and more reliable and robust technologies are needed. To address this issue and to evaluate the predictive value of HER2 on chemotherapy, we investigated a cohort of 278 patients from the GeparTrio trial, a prospective neoadjuvant anthracycline/taxane-based multicenter study. In the GeparTrio trial, patients were not treated with any anti-HER2 therapy, as this was not standard therapy at this time. The HER2 status was analyzed by three different approaches: local and central evaluation using immunohistochemistry combined with in situ-hybridization as well as evaluation of HER2 mRNA expression using kinetic RT-PCR from formalin-fixed, paraffin-embedded (FFPE) tissue samples using a predefined cutoff. HER2 overexpression/amplification was observed in 37.3% (91/244) and 17.9% (41/229) of the informative samples in the local and central evaluations, respectively. Positive HER2 mRNA levels were found in 19.8% (55/278). We observed a highly significant correlation between central HER2 expression and HER2 status measured by kinetic RT-PCR (r = 0.856, P < 0.0001) and an overall agreement of 95.6% (κ statistic, 0.862, CI 0.77–0.94). Further, central HER2 as well as HER2 mRNA expression were predictors for a pathological complete response after neoadjuvant anthracycline/taxane-based primary chemotherapy in a univariate binary logistic regression analysis (OR 3.29, P = 0.002; OR 2.65, P = 0.004). The predictive value could be confirmed for the central HER2 status by multivariate analysis (OR 3.04, P = 0.027). The locally assessed HER2 status was not predictive of response to chemotherapy. Our results suggest that standardized methods are preferable for evaluation of HER2 status. The kinetic RT-PCR from FFPE tissue might be an additional approach for assessment of this important prognostic and predictive parameter but has to be confirmed by other studies.