It’s the latest hope offered to many patients at major cancer research centers: Let us sequence your tumor, and maybe we can match it to a drug that will beat back the malignancy. But the reality is that genomic analysis still only helps a small fraction of cancer patients. “More often than not, we don’t know what to do with the information,” says Charles Sawyers, a researcher at Memorial Sloan Kettering Cancer Center (MSKCC) in New York City.
Now, several large U.S. and European academic medical centers are hoping to change that by pooling data on patients’ tumor genomes and their clinical outcomes. The project, announced today by the American Association for Cancer Research (AACR) at a meeting in Boston, has a mouthful of an acronym: GENIE, which stands for Genomics, Evidence, Neoplasia, Information, Exchange. It is spearheaded by Sawyers, who helped develop the leukemia drug Gleevec, one of the most successful gene-targeted therapies.
Sawyers says that GENIE grew out of conversations with colleagues at other cancer centers who are amassing tumor genetic data at a “mind boggling” rate. “It dawned on us that despite our history of tending to be silos, we could actually benefit from a pooled data sharing operation,” he says. So far, seven research centers have signed on, including MSKCC, the Dana-Farber Cancer Institute in Boston, Johns Hopkins University in Baltimore, Maryland, and three organizations in Canada and Europe. AACR has put up $2 million for a 2-year pilot stage to coordinate and store the data.
The centers tend to offer sequencing to late-stage cancer patients whose cancer has spread. Usually they aren’t sequencing entire tumor genomes; instead, they sequence a set of known cancer genes, such as BRAF, which is often mutated in melanomas, and EGFR, which drives the growth of many lung tumors. The sequencing report then typically goes to a board of experts who determines whether it points to a specific treatment. But although BRAF and EGFR mutations can be targeted by drugs, often the tumor has only rare mutations whose role in cancer isn’t well understood.
In other cases, says Sawyer, a seldom-seen mutation suggests a “really interesting, tantalizing hypothesis that is being acted on” at one research center, by giving the patient a drug that had been approved for another cancer. The problem, he explains, is that “the outcomes of those independent single patient decisions aren't captured in an organized way.”
GENIE will allow researchers to share these rare treatment cases and reach the numbers they need to draw firm conclusions. The project could also reveal mutations that influence a patient’s prognosis.
Working out legal agreements with each center for securely sharing data from medical records wasn’t an easy task, Sawyers says. Leaders also decided not to develop a standard gene panel and instead let each center keep using its own, sometimes custom-made set. They vary in size—MSKCC’s is one of the biggest with 410 genes—but most lists overlap, Sawyers says. All sequencing data will come from labs that meet quality standards needed to use the data in clinical care.
Another challenge was figuring out how to give researchers who are sharing their hard-earned data a chance to publish any significant discoveries. GENIE members will get 3 months after sequencing a patient’s tumor to submit the data to Sage Bionetworks, a nonprofit organization in Seattle, Washington, that is compiling it in a database. For the next 6 months, only the contributing institution can see that individual patient’s record within the database. For the subsequent 6 months, it will be open to the full consortium. Finally, the patient’s data will be available to broader research community.
So far, the registry has data on 17,000 patients and hopes to reach 100,000 within 5 years. And Sawyers hopes it will only grow. “It’s sort of sociology experiment for institutional collaboration,” he says.