RESUMO
AIMS: To compare the sensitivity and specificity of new rabbit monoclonal antibody SP3 with those of mouse monoclonal and rabbit polyclonal antibodies using HER2 amplification defined by chromogenic in situ hybridisation (CISH) as the gold standard. METHODS: Serial sections from tissue microarrays (TMAs) containing 84 breast carcinomas were submitted to CISH (Zymed HER2 Spot-Light kit) and immunohistochemistry, using NeoMarkers SP3 (rabbit monoclonal), DAKO A0485 and DAKO HercepTest (polyclonal), Novocastra NCL-CB11, Cell Marque CM-CB11, and Genentech 4D5 (mouse monoclonal). RESULTS: The best antibody concordance was between SP3 and HercepTest (kappa = 0.74). SP3, A0485 and HercepTest detected all HER2 amplified tumours, but were less specific than mouse monoclonal antibodies. 3/38 (7.9%) and 8/38 (21.0%) non-amplified tumours were scored as 3+ using SP3 and A0485, respectively. 3/46 (6.5%) amplified tumours were negative for NCL-CB11. SP3, HercepTest and A0485 showed no gene amplification on 55%, 62.5% and 92.3% of the 2+ scored tumours, but most of the 2+ scored tumours using monoclonal antibodies were amplified by CISH (80-92.3%). CONCLUSIONS: SP3 is more sensitive than mouse monoclonal antibodies for Her2 assessment. However, HercepTest, CB11 and 4D5 show higher specificity than SP3 for the identification of HER2 gene amplification. Mouse monoclonal antibodies show less Her2 2+ tumours; most are amplified by CISH.