Taming the Tumor Genetics Revolution
Daily, some patients’ cancer care is being revolutionized by Next Generation Sequencing (NGS). Clinicians suddenly wielding the magical ability to scan patient tumors for thousands of mutations at a pop—instead of one or two—are finding many tumors’ underlying “biomarkers” are vastly different than thought.
One example: some patients’ bladder cancers look like breast cancers. Clinicians seeing this can take completely different meds off the shelf—breast not bladder cancer meds--and take far better aim at disease.
The hope is NGS is revolutionizing oncology. But the reality right now, says tumor biomarker expert Daniel Hayes, is “chaos. It is very Wild West out there right now.”
The problem: the system for establishing biomarkers in the first place is disorganized, inefficient, non-centralized, inadequately reimbursed, and woefully underfunded, Hayes and a blue ribbon panel reported in a recent Science Translational Medicine article. The group is calling for industry and academia to recognize this, and begin conducting “high level meetings” to rectify it. “I am sure we will get pushback in all fields,” says Hayes, co-chair of the American Society of Clinical Oncology’s Expert Panel for Tumor Marker Practice Guidelines.
“But this is absolutely critical.”
Via the Cancer Genome Atlas (TCGA) project, the government is spending $100 million a year to do something extraordinary: take advantage of the tumbling cost of genome sequencing to fully sequence patient tumors in many different cancers. (And the cost truly is spiraling. The first human genome was sequenced in 2003 for $3 billion, taking 13 years. Now it takes less than a day, costing less than $10,000.)
This ambitious new drive to sequence all or most mutations in given tumors has resulted in almost quarterly papers, pumped out by huge TCGA teams, offering new cancer insights—and new potential biomarkers. Thus, as some TCGA researchers were finding some bladder cancers express the breast cancer biomarker HER2—letting such patients take the successful drug Herceptin (see earlier Bioscience Technology article)—others were finding that other breast cancers look more like ovarian cancers.
“In tumor genetics, there is no question that the introduction of NGS technologies has already added substantially to the care of individual patients with cancer, as therapies aimed at genomic targets offer the promise of improved efficacy and decreased toxicity,” Kenneth Offit, chief of the clinical genetics service at Memorial Sloan–Kettering Cancer Center, told the Journal of the National Cancer Institute this summer.
That article cited another example of NGS wizardry: one man’s pancreatic neuroendocrine cancer turned out to resemble, genetically, gastrointestinal cancer. So he was given Gleevec for the latter cancer and “is alive and well two years later when he would have been dead otherwise,” Fox Chase Cancer Center vice president Jeff Boyd said.
Fox Chase uses a test that looks at 45 genes. This summer, Massachusetts General Hospital will start looking at 1,000 genes at a time. The Dana Farber Cancer Institute is now, via a program called Profile, testing all cancer patients for a few hundred data points. Greenville Hospital System University Medical Center is starting a program.
“That’s all exciting,” says Hayes. “But there is an important distinction between a biomarker and an analytically validated test for the biomarker that has been shown, with high levels of evidence, to improve patient care. For example, HER2 is a biomarker. It can be measured in tissue or blood. One can determine abnormalities in DNA, RNA, and protein, using many different assays for each. Some of these have excellent analytical validity. A few have high levels of evidence for clinical utility and are strongly recommended for routine clinical use.”
But it takes a long time to get there, he says. Too many biomarkers, like the PSA prostate cancer test, disappoint. “Right now the TCGA is in exploratory mode to find things that might be useful. But the hard work comes next. Are there highly accurate assays for the presumed important mutations or chromosomal copy number variations? And do they actually predict benefit from the presumed drug, as demonstrated by scientifically designed clinical trials?”
University of Pittsburgh Medical Center oncologist Michael Lotze cautions that over-regulation kills. “We don’t want higher hurdles for biomarkers. One reason we have so few biomarkers is that much of the complex landscape of cancer is dictated by rounds of tumor cell growth and death. One can find much of the flotsam and jetsam in the serum.”
Hayes agrees. “I am a capitalist, not a communist. This will take delicate balancing. We can’t demand higher reimbursement for biomarker tests if the biomarkers aren’t better, for one example. Many areas must be addressed simultaneously. Biomarkers are as important as drugs.”