Agreement of Different Methods for Tissue Based Detection of HER2 Signal in Invasive Breast Cancer

Abstract

Breast cancer is the second leading cause of cancer mortality amongst American women. The HER2 gene encodes a cell surface receptor that affects cell proliferation and has been recognized as a diagnostic factor in treatment selection for invasive breast cancer. Examine accuracy in HER2 detection between manual count, computer assisted, and automated tiling algorithm. 42 randomly selected invasive breast cancer specimens were enumerated by fluorescence in situ hybridization (FISH)for HER2 and CEP17 markers using the Vysis HER2 assay (AbbotLaboratory, North Chicago, IL). Specimens were tested using three methods: Manual, computer assisted nuclei selection (Tissue FISH MetaSystems, Newton, MA), and automated enumeration (MetaSystems, Newton, MA). The greatest bias and widest agreement limits for HER2 and CEP17 were seen in Automatic versus Manual, the gold standard. HER2 values greater than 6 possessed the greatest bias and widest agreement limits. CEP17 comparison showed similar bias and agreement limits for each comparison. Kappa values indicated good agreement for all methods although Tissue FISH and Manual possessed better agreement. Higher agreement at lower HER2 & CEP17 count maybe due to fewer chromosomal aberrations, in which selection of field of views has less variation between methods. Alternatively, increased background signals seen in polyploidy may be responsible for the variations in signal count. Manual and Tissue FISH demonstrated good agreement amongst by both Altman Bland and Cohen's Kappa. While the automatic method has good agreement at lower HER2, the sharp increase in variability at higher HER2 counts illustrates a limitation of the automatic method.

Keywords: Altman bland; Breast neoplasms/diagnosis; Breast neoplasms/treatment; Fish; HER2 Recptor.