Validation of breast cancer biomarkers between a field immunohistochemistry laboratory in Nigeria and its US-based counterpart.

Abstract Abstract #4039 Introduction: The importance of hormone receptor status in assigning treatment and the potential use of HER2 targeted therapy have made it imperative for laboratories to improve detection techniques. As inter-laboratory variability in immunohistochemical (IHC) tests may also affect epidemiologic studies of breast cancer subtypes in different countries, we conducted a validation study of breast cancer biomarkers between a well-established laboratory in the US and a field laboratory at the Institute for Medical Research and Training at the University College Hospital in Ibadan Nigeria.
 Method: 232 breast tumor blocks were evaluated for ER, PR, and HER2 at both laboratories using tissue micro arrays (TMA) technique. Web-based conferences were held periodically to discuss IHC staining protocols, standardize scoring systems and to resolve discrepant cases. Pathologists used whole slide imaging for joint review and kappa statistic (κ) was used to indicate concordance between the two laboratories. Fluorescence in situ hybridization was carried out to confirm HER2 status in all cases.
 Results: Initially, concordance analysis revealed an agreement of 91% (κ=0.52) for ER, 85% (κ=0.49) for PR, and 80% (κ=0.39) for HER2 between the two labs. Antigen retrieval techniques and scoring methods were identified as important reasons for discrepancy. After quality assurance and training, the agreement improved to 92% (κ=0.53) for ER, 88% (κ=0.64) for PR, and 94% (κ=0.75) for HER2. To date, florescence in situ hybridization (FISH) has been completed for 67 cases to confirm HER2 status, out of which 16 (24%) were shown to amplify the HER2 gene, 6 out of the 12 discordant HER2 results were resolved by FISH.
 Conclusion: We found web-based conference with TMA and digital microscopy a useful and cost-effective tool for quality assurance of IHC, consultation and collaboration between distant laboratories. Quality improvement exercises in testing of tumor biomarkers will reduce misclassification in epidemiologic study of breast cancer subtypes and provide much needed capacity building in resource poor field sites.
 Acknowledgement
 This study was carried out with the support of Breast SPORE NCI P50 CA125183, Breast Cancer Research Foundation and the Lee Jeans Entertainment Industry Fund. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 4039.