Abstract 3143: Correlation of HER2 expression by IHC, DNA Microarray and FISH in breast cancer

Abstract

Abstract HER2 overexpression occurs in approximately 15-20% of patients with breast cancer and is associated with aggressive disease and decreased survival. HER2 status is predictive of response to trastuzumab and lapatinib. Given the importance of HER2 positive disease, accurate evaluation of HER2 status is essential. The aim of this study is to provide insights into the relationship between HER2 expression by immunohistochemistry, DNA microarray and FISH in a large cohort of 1,032 Breast cancers. HER2 protein expression was determined using antibody clones (4B5) and interpreted per ASCO/CAP scoring criteria. Samples scored as equivocal (>2+;> 10% to <=3+; <=30%) were required to undergo further assessment with FISH HER2 testing(Pathvysion). All samples were tested for HER2 by DNA microarray provided there was sufficient quantity of RNA in all tumor samples. Investigation of two-way comparisons using qualitative (categorial) outcome measure (overexpressed, underexpressed, equivocal) yielded statistically significant results but demonstrated poor kappa values, showing only moderate to slight agreement between the assays (IHC/FISH K=.096, p<.001; FISH/Microarray K=.079, p<.05; IHC/Microarray K=.126, p<.001). In-depth analyses indicate that equivocal or negative protein expression trend towards non-amplification (76.9%, n=837 and 90.5% n=182, respectively, p< .001). Low RNA levels are most often associated with IHC negativity (61.2%, n=762) but are also noted in samples with HER2 overexpression (38.4%, n=478). The least structured relationship was observed between DNA Microarray and FISH. Pearson correlation coefficients, based on intensity of staining, Quick score (IHC), gene ratio (DNA Microarray) and HER2-FISH ratio indicate that both IHC Quick score and intensity of staining exhibit modest correlations (r=.529, n=979, p<.001; r=.486, n=981, p<.001, respectively). This is well evidenced in the association between HER2 protein expression and median HER2-FISH ratio, which is 5.32 (SD=2.47) in overexpressed and 1.06 (SD=.34) in underexpressed samples, even when controlling for equivocal results (Kruskal Wallis <.001, n=142). DNA Microarray showed the least concordance with the other assys (FISH r=.440, n=573, p<.001), IHC/ Quick score r=.377, n=573, p<.001; IHC/intensity r=.342, n=573, p<.001). To conclude, our data suggest that – due to poor inter-assay agreement – Breast cancers with HER2 IHC and HER2 FISH equivocal results cannot be resolved by DNA Microarray. Protein expression by IHC is the most robust and cost effective way to test for HER2 expression when performed per ASCO/CAP guidelines and it correlates well with gene amplification by FISH. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 3143. doi:10.1158/1538-7445.AM2011-3143