In 2006, the NIH launched a $300 million initiative called The Cancer Genome Atlas to create a dataset of genomic changes across an array of cancers. Before the program was launched, researchers tended to study and treat cancers in isolation. Breast cancer was separate from stomach cancer, which was separate from prostate cancer, and so on. Over the past decade, however, hundreds of scientists participated in the NIH program to gather information on 11,000 tumors that span 33 types of cancer. Their findings, published in 27 journal articles this month, represent the most comprehensive cross-cancer research to date.
The studies, collectively called the Pan-Cancer Atlas, include insights like gene alterations that are shared across cancer types and may help inform future research.
“The scale of the dataset is massive. It’s just the largest and most complete dataset out there, period,” says Pan-Cancer Atlas collaborator Nikolaus Schultz, a cancer biologist at Memorial Sloan Kettering Cancer Center.
To gather material for the project, tumor tissue was acquired from consenting patients and sent to one of three processing laboratories called Biospecimen Core Resources. Different teams then assessed specific characteristics of the samples, such as DNA mutations, structural DNA changes, or alterations in how RNA is transcribed to express proteins — alterations that turn healthy cells into carcinogenic ones.
The Pan-Cancer Atlas authors mined the trove of data to uncover insights and trends. The team compared, contrasted, and clustered tumors to investigate traits more relevant than just the tissue in which the tumor arose. For example, they discovered that some DNA changes, gene expression, and chromosome numbers in tumor cells were similar for tumors across different organs. Tumors composed of a specific type of cell, called a squamous cell, had especially similar characteristics, says the report’s lead author Katherine Hoadley, an assistant professor of cancer genetics at the University of North Carolina.
The investigators also presented three processes that fuel cancer development, assessed alterations in the pathways that direct cell growth, and identified which subtypes are more likely to respond to immunotherapy.
Understanding the similarities and differences between cancer subtypes provides a foundation for developing new treatments. “Cancer is not a single disease. It’s 300 different diseases. Each one of them will require a cure,” says Jean Claude Zenklusen, director of the The Cancer Genome Atlas. Having organized data for each subtype will help scientists tailor treatments for each one, he says.
The data may also help patients if their specific cancer can be classified based on both its origin and its molecular makeup. “The molecular information can help us understand better who’s going to benefit from certain treatments and who’s not. That’s the precision medicine concept,” says Carolyn Hutter, the director of the division of genome sciences at the National Human Genome Research Institute.
Hutter hopes that in the future, scientists will review the Pan-Cancer Atlas and use it as a jumping off point for new research and clinical studies. Investigators can use the Atlas to efficiently identify a mutation, determine its prevalence across cancers, and study how patients with that mutation fared, for example.
“Once upon a time, that required going to a dozen separate places, collecting a dozen separate data sets, and putting them together yourself. Now all that data is organized,” says Leslie Cope, an associate professor of oncology at Johns Hopkins Medicine. The Pan-Cancer Atlas makes it easy to step back and see the big picture, he says.