Whole-genome sequencing may be one of the best bets in the race toward personalized medicine, especially on the oncology track.
Researchers finished sequencing the human genome less than a decade ago, at a cost of $3 billion. Clinicians are now deciphering the genomes of individual tumors to guide therapeutic decisions, and envisioning a not-so-distant future when the technique might be no more unusual than a blood test.
True believers are positioning whole-genome sequencing as one of the best bets in the race toward personalized medicine, especially on the oncology track. However, to get there, clinical researchers must clear additional financial, logistical, and ethical hurdles.
Multiple laboratories have already lowered the turnaround time for a completed tumor genome to a few weeks or less, and a $1000 genome will soon be within reach. Edward Abrahams, PhD, president of the Washington, DC-based Personalized Medicine Coalition, describes the declining cost as a major advancement in bringing the notion of truly personalized medicine to the brink of possibility. “It's really hard not to be excited about the opportunity,” he says. At the same time, he concedes, the clinical usefulness of the approach has yet to be fully demonstrated. “We've had glimpses of the future.”
Those peeks, however, have been enough to entice an entire industry to scale up its efforts to promote whole-genome sequencing, and Dr. Abrahams says it is just a matter of time before the remaining obstacles fall away. As Eric Topol, MD, director of the Scripps Translational Science Institute in La Jolla, California, explains in his book, The Creative Destruction of Medicine, cancer is a genomic disease whose care is among the most likely to be transformed by next-generation sequencing.1
A major goal of cancer genomics is to identify “actionable” mutations that drive a tumor and can be targeted with available therapy, and the ever-lower costs of sequencing technology have brought that goal closer to clinical practicality.
However, there is a catch. “Then you have a 5-inch-thick set of papers on your desk for the bioinformatics. That's where the cost is,” says Gail H. Vance, MD, professor of medical and molecular genetics and pathology and laboratory medicine and director of the division of diagnostic genomics at Indiana University School of Medicine in Indianapolis. In turn, those data points must be distilled into a clinically relevant interpretation whose added expense may never be recouped. And the entire process might need to be repeated several times to capture the evolution of a cancer that recurs or is resistant to therapy.
Building the Case for Clinical Genomics
Despite the challenges, many researchers are eager to build the case for whole-genome sequencing as more informative and financially feasible, especially as additional genes become implicated in specific cancers. “We're going to do the whole genome because it's more cost-effective to do the whole thing rather than do it 1 gene at a time,” says Mark S. Boguski, MD, PhD, associate professor of pathology at Harvard Medical School in Boston, Massachusetts, and cofounder of Genome Health Solutions Inc. “But we're going to restrict our interpretation to the clinical question at hand.”
That approach is fundamentally different from the predictive genomics that have reigned since the completion of the Human Genome Project. Dr. Boguski says precision diagnostics to address an acute medical problem will drive personalized medicine far more than theoretical probabilities of disease development based on germline DNA. The Cancer Genome Atlas has provided several tantalizing hints of where better genome-aided diagnostics might lead. To date, the project has released comprehensive genomic analyses of 3 cancer types, including a recent study published in Nature that scrutinized colon and rectal tumors from 224 patients and sequenced the entire genomes of 97 of those tumors.2
Among the 24 genes that harbored significant tumorlinked mutations, the study pointed out several unexpected targets. For example, approximately 5% of the colorectal tumors analyzed contained multiple copies of the ERBB2 oncogene. The anomaly could potentially open up a new patient pool for the anti-ERBB2 drug trastuzumab. “The challenge there is, if we find something like that, is the oncologist willing to consider treating off-label?” Dr. Boguski asks. “Are insurance companies willing to pay for off-label use of those targeted therapies?”
Beyond the collective uncovering of new drug targets, researchers have begun publishing case studies affirming the new technology's potential for individual patients. Elaine Mardis, PhD, codirector of the Genome Institute at Washington University in St. Louis, Missouri, and her collaborators made headlines this past July when a whole-genome sequence and RNA analysis radically improved the prognosis of a research colleague who had been diagnosed with adult acute lymphoblastic leukemia (ALL) and experienced disease recurrence twice.
Genomic analysis of his tumor cells revealed a highly overexpressed fms-related tyrosine kinase 3 (FLT3) gene, a driver mutation not usually associated with ALL but one that was likely behind the uncontrolled proliferation. Physicians were able to target the FLT3 overexpression with sunitinib, a drug approved by the US Food and Drug Administration for the treatment of patients with advanced renal cell carcinoma and other tumors. Once in remission, the patient underwent his second bone marrow transplant, which appears to have been successful to date.
Given the publicity surrounding such successes, Dr. Boguski believes patient demand will put increasing pressure on physicians to offer whole-genome sequencing, and then on insurance companies to pay for it. Since the Washington University ALL case went public, a medical oncologist there told Dr. Boguski that approximately 1 in 5 patients coming into his leukemia clinic have asked if they can have their tumor genome analyzed. “We don't want that patient demand to result in something that's not sensible, so we're very interested in building the case through comparative effectiveness research that this can lead to cost avoidance,” Dr. Boguski says. “The way to think about this is not how much the tests cost up front, but how much cost it allows you to avoid after the diagnosis by eliminating the need for trial-and-error empiric therapies.”
Unforeseen Results of Tumor Genomic Analysis
However, whole-genome sequencing also has spurred ethical dilemmas over what researchers are calling “unintentional findings” or “incidentals”: discoveries such as germline mutations associated with disease predisposition. A 2011 report by Dr. Mardis and her colleagues describes a patient whose bone marrow and healthy skin genomes were sequenced several years after her death as part of a project to understand why some patients initially treated for solid tumors subsequently developed therapy-related acute myeloid leukemia.3
During the course of the sequencing, the genomic researchers discovered a huge germline deletion within 1 copy of her tumor protein 53 (TP53) gene encoding the P53 tumor suppressor protein, meaning that she had the exceedingly rare Li-Fraumeni syndrome. A more focused polymerase chain reaction–based test had not caught the deletion. The patient had subsequently lost the remaining wild-type TP53 allele, and been diagnosed with breast cancer and ovarian cancer, before succumbing to therapy-related acute myeloid leukemia. None of her children had any idea that they each had a 1 in 4 chance of inheriting the same Li-Fraumeni susceptibility, with its greatly elevated cancer risk, until a genetic counselor consulted by the researchers contacted the family to obtain consent and then inform them about the mutation.
Dr. Mardis says the case emphasizes the importance of informed consent and notification of patients' families. “If I were one of the kids of this woman, I would sure as hell want to know about it, because it shapes the rest of your life going forward, and I'd rather be knowledgeable about it than ignorant of it,” she says.
Such ethical considerations, Dr. Vance says, must be worked out before the testing is first offered. “That's key here, because otherwise you drop a bomb on someone who, first of all, has a disease, and you're telling them what their mutations are and how you're going to direct the therapy, and, ‘Oh, by the way, you have this too.’” Many clinical laboratories are moving toward a policy of reporting only those gene mutations that are relevant to an acute medical condition, due in part to the fear of legal liability.
Regardless of how laboratories handle the reporting issue, researchers say the clinical approach will require more interaction between patients and physicians to ensure that, as part of the quickened pace toward personalized medicine, patients understand what will be explored, what will be left out, and why.
BRYN NELSON IS A FREELANCE MEDICAL JOURNALIST.
Cytosource Reader Poll #8:
Q: What is the biggest obstacle facing wholegenome sequencing's bid to become a viable clinical service?
A. Upfront costs and insurance reimbursement.
B. Demonstration of clinical usefulness.
C. Bioinformatics and interpretation logistics.
D. Ethical and legal challenges.
Take the poll online at www.cancercytojournal.com.
The results will be published in the February 25, 2013 issue.
JUNE POLL RESULTS
Q: Within the next decade, cancer drug discovery based on epigenetics:
38% Will not live up to expectations.
0% Will be useful mainly for sensitizing cancers to existing drugs.
0% Will be beneficial primarily for blood-borne tumors.