Cancer: Identifying the best ovarian cancer models

Some of the most popular ovarian cancer cell lines used for research may not be good models of ovarian cancer, a genomic analysis published in Nature Communications reveals.

Cancer cell lines derived from clinical tumour samples are frequently used in cancer research and have led to many new insights into cancer biology. Nikolaus Schultz and his team analysed publicly available datasets from two large-scale genetic sequencing projects - the Cancer Genome Atlas, which defined the genomic features of clinical ovarian cancer samples, and the Cancer Cell Line Encyclopedia, which contains similar data for about 1,000 cancer cell lines used for research. Schultz’ team matched these data against each other, ranking about 50 ovarian cancer cell lines by their genomic similarity with clinical samples from patients with high-grade serous ovarian cancer (HGSOC). They find that the two most popular ovarian cancer cell lines, accounting for 60% of all published research articles on HGSOC, do not resemble HGSOC very well, while the 12 most suitable ovarian cancer cell lines were used in only 1% of publications.

The study provides a methodological framework that could also be applied to other cell lines and different types of cancers in order to assess their validity as research tools.