Challenges in array comparative genomic hybridization for the analysis of cancer samples

PURPOSE: To address some of the challenges facing the incorporation of array comparative genomic hybridization technology as a clinical tool, including archived tumor tissue, tumor heterogeneity, DNA quality and quantity, and array comparative genomic hybridization platform selection and performance. METHODS: Experiments were designed to assess the impact of DNA source (e.g., archival material), quantity, and amplification on array comparative genomic hybridization results. Two microdissection methods were used to isolate tumor cells to minimize heterogeneity. These data and other data sets were used in a further performance comparison of two commonly used array comparative genomic hybridization platforms: bacterial artificial chromosome (Roswell Park Cancer Institute) and oligonucleotide (Agilent Technologies, Santa Clara, CA). RESULTS: Array comparative genomic hybridization data from as few as 100 formalin-fixed, paraffin-embedded cells isolated by laser capture microdissection and amplified were remarkably similar to array comparative genomic hybridization copy number alterations detected in the bulk (unamplified) population. Manual microdissection from frozen sections provided a rapid and inexpensive means to isolate tumor from adjacent DNA for amplification and array comparative genomic hybridization. Whole genome amplification introduced no appreciable allele bias on array comparative genomic hybridization. The array comparative genomic hybridization results provided by the bacterial artificial chromosome and Agilent platforms were concordant in general, but bacterial artificial chromosome array comparative genomic hybridization showed far fewer outliers and overall less technical noise, which could adversely affect the statistical interpretation of the data. CONCLUSIONS: This study demonstrates that copy number alterations can be robustly and reproducibly detected by array comparative genomic hybridization in DNA isolated from challenging tumor types and sources, including archival materials, low DNA yield, and heterogeneous tissues. Furthermore, bacterial artificial chromosome array comparative genomic hybridization offers the advantage over the Agilent oligonucleotide platform of presenting fewer outliers, which could affect data interpretation.