The demand for both reflexed and primary fluorescence in-situ hybridization (FISH) testing in the clinical setting is increasing. Relevant literature has reported the incidence of HER2 overexpression in 20% to 30% of cases, but some reports suggest that HER2 gene amplification rates are substantially lower. Published data, however, on primary FISH assessment from a single institution is limited, especially information about the frequency of the anomalous genotypes defined by FISH. We report our experience with primary FISH testing in 742 consecutive cases of breast cancer, in the calendar year 2006. Eighty percent (595/742) of the breast cancer cases were not amplified for HER2 (HER2/CEP17=0.8-1.9), whereas 19% (142/742) of cases were HER2 amplified (HER2/CEP17>or=2.0). Among the HER2-amplified cases, 3% (19/742) were low-level amplified (HER2/CEP17 ratio=2.0-2.5). Genotypic heterogeneity, defined as >5% but <50% of the tumor cells demonstrating HER2 gene amplification, was observed in 5% (40/7242) of the cases. HER2 monoallelic deletion (HER2/CEP17<or=0.7) was demonstrated in 2% (12/742) of the cases and CEP17 monosomy (1 CEP17 signal in >80% of tumor cells) was observed in 2% (13/742). Polysomy, if defined as CEP17 spot count 3.0 or more in at least 80% of tumor cells, was observed in 3% (20/742) of the cases. These data may be helpful as benchmarks for other institutions initiating primary FISH analysis for HER2 genotyping.