Clinical laboratory testing for HER-2/neu gene amplification by fluorescence in situ hybridization is not widely used in diagnostic pathology laboratories. A bright field alternative permitting direct visualization of gene amplification using conventional microscopy may be more readily incorporated into routine diagnostic pathology practice. Interobserver reproducibility represents an important component of the validation of such an assay. We tested the hypothesis that a first-generation bright field alternative to fluorescence in situ hybridization, a Nanogold (Nanoprobes, Inc, Yaphank, NY, USA) (or gold-label)/autometallographic assay for HER-2/neu gene amplification in breast carcinoma, can be reproducibly interpreted by pathologists. Reference standard was direct fluorescence in situ hybridization supplemented by RNA/RNA in situ hybridization. Reproducibility of selected conventional histologic parameters was captured based on a hematoxylin and eosin slide accompanying the GOLDFISH preparation (gold-facilitated autometallographic in situ hybridization) as an indication of comparative reproducibility. The average kappa among GOLDFISH observers was 0.84, which was at least or concordant of observers scoring nuclear grade (kappa = 0.50) and the presence of in situ carcinoma (kappa = 0.57) by conventional histopathology. The GOLDFISH assay was specifically designed for qualitative interpretation, thus obviating the need for oil immersion microscopy and signal enumeration, and its interpretation was highly reproducible among five pathologists.