The HER2 gene is an important prognostic and therapeutic marker in newly diagnosed breast cancer. Currently, HER2 status is most frequently determined by immunohistochemical detection of HER2 protein expression on the cellular membrane surface or by fluorescence in situ hybridization analysis of HER2 gene copy number in fixed tissue using locus-specific probes for the HER2 gene and chromosome 17 centromere. However, these methods are problematic because of issues with intra- and inter-laboratory reproducibility and preanalytic variables, such as fixation time. In addition, the commonly used HER2/chromosome 17 ratio presumes that chromosome 17 polysomy is present when the centromere is amplified, even though analysis of the rest of the chromosome is not included in the assay. In this study, 97 frozen samples of invasive lobular and invasive ductal carcinoma, with known immunohistochemistry and fluorescence in situ hybridization results for HER2, were analyzed by comparative genomic hybridization to a commercially available bacterial artificial chromosome whole-genome array containing 99 probes targeted to chromosome 17 and the HER2/TOP2 amplicon. Results were 97% concordant for HER2 status, meeting the College of American Pathologists/American Society of Clinical Oncology's validation requirements for HER2 testing. Surprisingly, not a single case of complete polysomy 17 was detected even though multiple breast cancer cases showed clear polysomies of other chromosomes. We conclude that array comparative genomic hybridization is an accurate and objective DNA-based alternative for clinical evaluation of HER2 gene copy number, and that polysomy 17 is a rare event in breast cancer.