Evaluation of HER-2/<i>neu</i> Gene Amplification and Overexpression: Comparison of Frequently Used Assay Methods in a Molecularly Characterized Cohort of Breast Cancer Specimens

Michael F. Press(University of Southern California), Dennis J. Slamon(University of Southern California), Kerry Flom(University of Southern California), Jinha Park(University of Southern California), Jianyuan Zhou(University of Southern California), Leslie Bernstein(University of Southern California)
Journal of Clinical Oncology
July 15, 2002
Cited by 389

Abstract

PURPOSE: To compare and evaluate HER-2/neu clinical assay methods. MATERIALS AND METHODS: One hundred seventeen breast cancer specimens with known HER-2/neu amplification and overexpression status were assayed with four different immunohistochemical assays and two different fluorescence in situ hybridization (FISH) assays. RESULTS: The accuracy of the FISH assays for HER-2/neu gene amplification was high, 97.4% for the Vysis PathVision assay (Vysis, Inc, Downers Grove, IL) and 95.7% for the the Ventana INFORM assay (Ventana, Medical Systems, Inc, Tucson, AZ). The immunohistochemical assay with the highest accuracy for HER-2/neu overexpression was obtained with R60 polyclonal antibody (96.6%), followed by immunohistochemical assays performed with 10H8 monoclonal antibody (95.7%), the Ventana CB11 monoclonal antibody (89.7%), and the DAKO HercepTest (88.9%; Dako, Corp, Carpinteria, CA). Only the sensitivities, and therefore, overall accuracy, of the DAKO Herceptest and Ventana CB11 immunohistochemical assays were significantly different from the more sensitive FISH assay. CONCLUSION: Based on these findings, the FISH assays were highly accurate, with immunohistochemical assays performed with R60 and 10H8 nearly as accurate. The DAKO HercepTest and the Ventana CB11 immunohistochemical assay were statistically significantly different from the Vysis FISH assay in evaluating these previously molecularly characterized breast cancer specimens.


Related Papers

No related papers found

Powered by citation graph analysis