Researchers Design Innovative System to Universally Catalogue Cancer Gene Variants

Posted in GUMC Stories

December 9, 2016 – A patient diagnosed with a tumor subtype previously unseen by their oncologist may feel concerned that, because there aren’t many people with her same diagnosis, drug companies have not developed tailored therapies and less treatment guidance exists than for more common cancers.

However, that patient’s cancer may be more common than their oncologist believes, says Subha Madhavan, PhD (new window), director of Georgetown Lombardi’s Innovation Center for Biomedical Informatics (ICBI) (new window). The issue, she says, is that data sharing in the cancer community is still in its infancy. While some data sets haven’t been shared, others have been shared in different formats, making it difficult to compare cancers and identify patterns.

Even results of genetic tests that oncologists order for their patients, some of which include a panel of 400 genes, aren’t standardized.  “If we are really serious about precision medicine — individualized cancer care — then we all have to be on the same page when discussing molecular landscape of patient tumors,” she says. “If all this information comes in a standard format, it will be easier for everyone to use.”

Data everyone can understand

In response to this issue, Madhavan and her colleagues developed a standard way for researchers everywhere to describe and catalog the gene mutations found in cancer biopsies. They also recently published a paper (new window) online in Genome Medicine offering ideas about how to fix this data streamlining issue, including a list of 18 key data elements.  

“This system will help researchers, clinicians and their patients,” says Madhavan, the senior author of the study. “It will improve the power of analysis by allowing researchers to group together treatment and outcomes data from patients with the same molecular cancer types, providing deeper insights into cancers subtypes; and it will result in a standardized molecular diagnostic report that is easier for patients to digest.”

Madhavan is a national leader in the effort to organize and utilize biomedical information efficiently and effectively. She co-chairs a multidisciplinary, multi-institutional group of researchers from the National Human Genome Research Institute-supported Clinical Genome Resource (ClinGen) (new window), a group that co-authored the study and includes researchers from Baylor College of Medicine, two of whom were co-lead researchers.

In the first weekend after the study appeared online, it was downloaded almost 500 times, demonstrating the desire for such standardization. “A lot of people are paying attention because this is the kind of guidance that the oncology research and clinical community has needed and wanted to represent their patient cancer molecular test data,” she says.

“Golden trove” of information

The information to be collected and standardized includes common information about a genetic change found in a tumor such as what chromosome it affects, whether it is within a gene and whether it was seen only in tumors or also in normal samples. In addition, each genetic change is characterized by how it might impact the cancer or treatment of the patient where it is found.

Standardizing this molecular information must also be done by existing databases that have used disparate systems to describe this content, Madhavan says.  There are many different teams of cancer researchers and clinical laboratories that provide information online and in other databases about gene variations associated with cancer. Two of the larger efforts to catalog cancer variation include Clinical Interpretations of Variants in Cancer (CIViC) (new window) and ClinVar (new window), a large database supported by the National Center for Biotechnology Information. Both of these groups are working with ClinGen to adopt MVLD (Minimal Variant Level Data) standards for data in their systems. MVLD is the standard the group developed to represent molecular data from patients.

If researchers adopt the proposed new system, cataloging efforts across databases will be unified, study authors say. Moreover, collecting this information in a standard form will result in new hypothesis generation and will almost certainly lead to new treatments, Madhavan adds.

“The discovery of variations in genes within tumor biopsies not only has been instrumental in understanding how cancer develops and spreads, but also how drugs can alter tumor growth. If we can virally build this standardized database of genetic involvement in all kinds of cancer, it makes complete sense that drug designers will mine this precious information to find new targets for more effective therapy.

“Data can be wasted or be a golden trove of insightful information,” she says. “Our goal is to make it the latter.”

Renee Twombly
GUMC Communications