The routine assessment of HER-2 expression can be affected by many immunohistological preanalytical and analytical variables. The evaluation of non-informative HER-2 tests, because of 2(+) scores, has been addressed in studies using in situ hybridization (fluorescent in situ hybridization (FISH) or chromogenic in situ hybridization (CISH)). There are very few studies that additionally checked 2(+) cases by quantitative reverse transcription-PCR (QRT-PCR). We analyzed totally 195 breast carcinoma cases, 70 of them showing 2(+) immunoreaction, with FISH/CISH and QRT-PCR. Confirmed amplification in 2(+) cases fell within the reported range (12.8% vs. 8-44%) and some of them showed lower mRNA levels indicating a genuine decrease of HER-2 protein as a mechanism for the non-informative score. In other cases, increased mRNA levels could be ascribed to HER-2 polysomy, verifying previous observations of immunohistologically detectable HER-2 polysomy. A remarkable subset of the 2(+) cases showed "normal" mRNA levels without amplification or polysomy and technical parameters as well as heterogeneity could be incriminated. The overall concordance of QRT-PCR and FISH was 93.8%, highest than most previously reported. Yet, the lack of clear cut-off mRNA values and the challenge of sample microdissection hinder QRT-PCR from claiming the status of a gold standard test for HER-2 evaluation.