Aims: HER2 protein is over-expressed in 15-30% of breast carcinomas. Immunohistochemistry (IHC) is a common and inexpensive method able to specifically detect HER2 protein. However, lack of standardization of IHC has been considered responsible for discrepancies in HER2 status assessment performed by IHC and fluorescence in-situ hybridization (FISH). This prompted us to perform a multicentric IHC calibration test to achieve a maximum accuracy of HER2-IHC compared with HER2-FISH taken as the reference method.
Methods and results: Twelve French laboratories participated in this study, including 119 cases of invasive breast carcinomas for which both fixed and frozen tissues were available. HER2 expression was determined in fixed tissues by individual in-house IHC techniques, using either CB11 (Novocastra, Newcastle, UK) or A0485 (Dako, Glostrup, Denmark) anti-HER2 antibodies. Two cut-off values were used: 10% and 60% of immunostained cells. In 116 of the 119 cases, HER2 gene status could also be determined by FISH on frozen sections, performed in a single laboratory. Results were centralized and compared. When suboptimal concordance between IHC and FISH was observed, IHC was calibrated and a second run was performed. The specificity, sensitivity and accuracy of IHC compared with FISH were noted before and after calibration. Forty-four out of 116 (38%) tumours showed HER2 gene amplification. Accuracy of IHC was complete in the first run for 6/12 laboratories. Calibration, necessary for the six others, relied mainly on the combination of a heat-induced epitope retrieval step with an increase of dilution of the primary antibody. In the second run, HER2 over-expression was found in 46 (40%) and 44 (38%) of the 116 cases, using 10% or 60% of stained cells as cut-offs, respectively. The corresponding accuracy rates were 93% and 95%.
Conclusions: This study showed that a high accuracy of IHC could be obtained for the determination of HER2 status in all laboratories using their in-house IHC technique, provided that a calibration process was performed. Antigen retrieval procedure, high dilutions of anti-HER2 antibody and the use of specific controls were crucial for HER2-IHC calibration. A 95% accuracy rate of IHC, using FISH as gold standard, was obtained by considering immunolabelling HER2-IHC results as a continuous variable, and taking 60% invasive stained cells as the cut-off for HER2 over-expression.