Copy number alterations in breast carcinomas: Comparison of different methods for assessment

Abstract

Breast cancer is a disease dominated by copy number alterations, but only HER2 copy number assessment analyzed by fluorescence in situ hybridization (FISH) is today included in the guidelines for breast cancer diagnostics. The anticipated inclusion of copy number estimation for additional DNA segments would be time consuming and cumbersome if analyzed with FISH. The development of Next Generation Sequencing (NGS) has opened the possibility of copy number assessment in multiple regions by targeted sequencing. Introduction of a panel aiming at detecting copy number alterations in breast cancer would need validation. We employed FISH analysis and designed and validated 17 bacterial artificial chromosomes (BAC) probes for selected chromosomal regions which are often subject to copy number alterations in breast cancer. Copy number analysis performed on formalin-fixes paraffin-embedded (FFPE) samples are subject to potential loss of chromosomal material due to sectioning artefacts. NGS uses bulk tumor samples and does not suffer from this effect. We therefore wanted to see if there is a systematic relationship in copy number difference between estimates using FFPE (i.e. sectioned cells) and estimates using whole cell imprints to identify a possible adjusting factor. Linear regression generated coefficients of large variation for the cell lines. We performed FISH analysis with the 17 probes on both imprints and FFPE sections of the breast cancer cell lines HCC1954 and HCC2218 and found that there were in general lower copy number estimates from FFPE sections. The average difference was about one copy number for the most heterogeneous cell line and somewhat less for the other, but there were large variations between the different regions in both cell lines so no specific adjusting factor was found. We then used the same validated probes in FISH analysis of FFPE tumor samples from 12 patients diagnosed with breast cancer and compared the copy number estimates with NGS estimates for the same regions. The copy number estimates by NGS largely aligned with estimates by FISH, but with certain notable discrepancies, reflecting the NGS test still being in a state of development. Further validation of the NGS test against the current gold standard for copy number assessment, FISH, combined with improved data analysis, is a crucial step preceding implementation in routine diagnostics.