Proficiency testing of immunohistochemical biomarker assays in breast cancer

Steroid hormone receptor expression and HER2 status have become an integral part of histopathologic characterization of breast cancer and corresponding biomarker assays have gained important prognostic and predictive impact. Because testing inaccuracy could provide a major hazard to modern breast cancer therapy, a laboratory proficiency testing program has been implemented in Germany using tissue microarrays (TMAs). In four consecutive annual trials with 142 laboratories participating on average per trial, estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (Her2) were determined immunohistochemically by participating laboratories followed by central review of all immunostains. Performance strongly depended on the ambiguity of expression of the target molecule in the test samples. In clearly positive (Allred score 7-8; Her2 3+) or negative tissue samples, the majority of participants (86%) achieved concordance rates exceeding 85%. By contrast, low expression of ER or PR (Allred score 3-4) as well as Her2 status 2+ led to considerable lower concordance rates ranging from 41% (Her2 2+) to 75% (PR). Poor reproducibility was predominantly due to inadequate laboratory performance whereas interobserver agreement (weighted kappa statistics) usually was high (> 0.81). Laboratories that participated in more than one of the four subsequent trials (n=110) showed a highly significant improvement of performance. In conclusion, a TMA-based proficiency testing of biomarkers in breast cancer has been implemented in Germany over a 5-year period and revealed reliable assessment of unambiguously positive and negative test samples. Low-expressing tumor samples with regard to steroid hormone receptor expression and Her2 status 2+ led to inaccurate evaluations by up to 59% of participants. Regularly participating laboratories showed a significant improvement of performance.