Although cancer researchers have made important discoveries with currently available cancer cell lines, several shortfalls limit the utility of these models. Cell lines don't exist for all cancer types, and don't capture the full genetic diversity of any cancer type. Two-dimensional models don't necessarily reflect the biology of three-dimensional tumours. And a lack of clinical data about how cancer-cell donors responded to treatment complicates interpretation of preclinical results.

To address these deficits, four US and European organizations are launching the Human Cancer Model Initiative (HCMI) to generate 1,000 new cancer models.

“This new project is timed perfectly to take advantage of the latest cell culture and genomic sequencing techniques to create models that are representative of patient tumours and are annotated with genomic and clinical information,” said Louis Staudt, director of the US National Cancer Institute (NCI)'s Center for Cancer Genomics, in a press statement. Contributing scientists will also use reprogrammed stem cells and organoids — 3D organ-like cell structures — to generate better models of disease. “This effort is a first step toward learning how to use these tools to design individualized treatments,” he said.

The NCI, the Sanger Institute in Cambridge, UK, Cancer Research UK and the Hubrecht Institute in Utrecht, the Netherlands, are funding and contributing to the HCMI.

The initiative will generate models for common cancers as well as for rare and childhood cancers. Cancer models, and associated data sets and annotations, will be shared with the research community.

If all goes to plan, the HCMI could eventually generate thousands more models as well. Researchers may ultimately need about 10,000 models to fully capture the diversity of relatively common genetic subtypes of cancer, Staudt told Nature News .