Discovering the Secrets of Brain Tumors
Physicians decide how to treat brain cancer in part based on what a patient’s tumor looks like under the microscope. But while the microscope can help classify a tumor, it cannot always reveal the molecular changes that drive the disease. And in an era of molecularly targeted therapies, this is what physicians and patients increasingly want to know.
Molecular markers have been identified and even genomic tests have been developed for breast and other common cancers, but progress has been slower in brain cancer. Obtaining the tumor tissue needed for molecular studies is difficult, and hundreds, if not thousands, of samples are needed to capture the genetic diversity of brain cancer.
Nonetheless, there is reason for optimism. Some of the first genomic studies of cancer included surveys of brain tumors, and for the first time, several clinical trials are enrolling patients based on tumor markers, such as combined deletions on chromosomes 1 and 19.
A database being developed by NCI and its partners around the country could also help move the field toward more individualized care. The Glioma Molecular Diagnostic Initiative (GMDI) is collecting and integrating molecular, genetic, and clinical data on hundreds of patients with gliomas, the most common type of brain tumor.
Changing the System
The March 1 issue of Cancer Research includes one of the first publications using the study’s database. The researchers propose a new classification system for gliomas based on tumor gene activity. While the results are considered preliminary, they illustrate the promise of an approach that seeks to integrate different types of information from patients with the disease, the researchers said.
“Unlike the standard classification system for gliomas, we came up with a system that is based purely on biology,” said lead investigator Dr. Howard A. Fine, chief of NCI’s Neuro-Oncology Branch in the Center for Cancer Research. “The results confirm and extend the findings of smaller glioma classification studies published previously,” he added.
Collectively, these studies show that there are subtypes of brain tumors defined not by their appearance under the microscope, but rather by the molecular pathways that are altered in tumors, said Dr. Mark Gilbert, professor of neuro-oncology at the University of Texas M.D. Anderson Cancer Center, who was not involved in the research.
“These studies all have the same goal,” Dr. Gilbert continued. “For a hundred years we’ve made treatment decisions on the basis of looking at brain tumor tissue under the microscope. Yet we know that two patients can have very different outcomes even though their tumors look identical, so it would be helpful if we could figure out the pathways that made the cancers different.”
Created to help attain that goal, GMDI has enrolled more than 800 patients with gliomas at 14 institutions around the country. A tumor sample collected during surgery for each patient is sent to NCI for molecular analysis. Clinical information on each patient is then linked to molecular and genetic descriptions of the tumor in a public database called REMBRANDT (Repository of Molecular Brain Neoplasia Data).
“We have built the largest cancer clinical genomics database for any tumor type, let alone an uncommon type such as glioma,” said Dr. Fine. In the Cancer Research study, the researchers used 159 gliomas from GMDI to develop an algorithm that assigns gliomas to six subtypes under two groups, glioblastomas and oligodendrogliomas. They validated the algorithm in three additional data sets, totaling nearly 700 gliomas.
The results confirm what physicians have always known—that some patients will do better than others for reasons that may not be understood, said Dr. Michael Prados, director of translational research in neuro-oncology at the University of California, San Francisco Comprehensive Cancer Center. He also leads a consortium that provided some of the annotated tumor samples used in the analysis.
“If we really want to change the outcomes for patients, then we need therapies that address changes in particular pathways,” added Dr. Prados. “We don’t have those therapies now, so the challenge going forward is: How will we use this information to make a difference in patients therapeutically?”
Exploring the Biology
As a next step, Dr. Fine’s group is characterizing the biology of the tumors to inform the development of new therapies. Until very recently, most patients with brain cancer received the same toxic drugs. But with a growing pipeline of molecularly targeted cancer drugs, physicians can start to evaluate these agents rationally in clinical trials.
Given that the most common subtypes may include only about 10 percent of all tumors, the selection of patients for these trials will be critical. A drug may perform poorly in a trial of unselected patients, and therefore could be discarded, even though it may be very effective for a specific subset of patients.
“For prognostic purposes, for drug development, and for the design of future clinical trials, we really need to be able to group brain tumors according to their underlying biology,” said Dr. Fine.
The genome surveys published last year and a recent follow-up report about relatively common mutations in brain tumors have set the stage for more individualized approaches to the disease. And while much more work remains to be done, everyone in the field seems to be moving in the same direction.
“Information is power, and these studies are all expanding the knowledge base,” said Dr. Prados. “This will let us develop therapies for patients in a more specific way.”
—Edward R. Winstead