March 23, 2011
Kevin Davies, Editor-in-Chief, Bio-IT World : Howard Jacob began his scientific career studying the genetics of rats with Broad Institute director Eric Lander, but for the past several years he has been telling anyone who will listen that genomic sequencing will reach the clinic by about 2014. His colleagues responded dismissively: “It will be a long time before we use genomic sequencing in the clinic—and no-one will ever pay for it,” they kept telling Jacob.
It turns out they were all wrong. In Milwaukee, Wisconsin, clinical sequencing has already arrived. Even more remarkably, insurance companies are (in some cases) offering reimbursement.
Jacob and his medical and informatics colleagues at the Medical College of Wisconsin made headlines late last year after they successfully identified the rogue gene mutated in a little boy named Nicholas Volker, who suffers from a rare and undiagnosed intestinal disorder. In a captivating talk at The Future of Genomic Medicine conference at the Scripps Institute last month, Jacob recalled the circumstances surrounding Volker’s diagnosis. (The Volker saga was reported in a superb three-part series in the Milwaukee-Wisconsin Journal Sentinel that has Pulitzer Prize written all over it.)
On June 27, 2009, Nicholas’ pediatric gastroenterologist, Alan Mayer, e-mailed Jacob to ask whether it was possible to sequence the genome of 6-year-old Volker. For six months, the boy underwent more than 100 surgeries to repair holes (fistulae) in his intestine, and his colon was removed piece by piece. Jacob decided his group could indeed sequence Volker’s genome—“How could you not?” he asked rhetorically—but elected to sequence just the exome (the protein-coding genes), as this would be simpler and, besides, it was all they could really afford at the time. Jacob collaborated with Roche/454 for five rounds of exome sequencing, at a cost of about $75,000.
Sequencing was easy compared to the data analysis problem—how to whittle down more than 16,000 unique DNA variants in Volker’s exome to a single causative mutation? Jacob paid tribute to his colleagues led by Elizabeth Worthey (ex-Rosetta), David Dimmock (ex-Baylor) and his bioinformatics team, even though their initial list of some 2,000 candidate genes proved incorrect.
The team found a point mutation in a highly conserved region of XIAP (X-linked Inactivation of Apoptosis). Jacob’s hunch is that Volker developed an immune response (perhaps to something he ate), but the mutated XIAP protein bound irreversibly, causing cell death in the intestines. But there was a more immediate concern: mutations in XIAP are linked with XLP2, a lymphoproliferative disease requiring a bone marrow transplant. Jacob informed the parents they had found a suspect mutation, but their son had a risk of cancer and could die if he didn’t have a transplant.
David Margolis, the program director of blood and bone marrow transplantation at the Children’s Hospital of Wisconsin, told Jacob: “If you put it in the medical record that it’s XIAP deficiency and the child is at risk for XLP2, I will do a bone marrow transplant.” “That’s the level of precision you have to have in order to effect an action,” said Jacob.
The Volker case study was published in Genetics in Medicine. Why such an obscure journal? “Because no-one else would take it!” Jacob shrugged. Editors and reviewers griped that Rick Lifton and colleagues at Yale had already reported a successful diagnosis using exome sequencing (in 2009), and that science is more than reporting a single case. I’m astounded at such short-sightedness.
No One-Hit Wonder
It would be tempting to dismiss Volker as a one-hit wonder. “We’re from the Medical College of Wisconsin. No-one’s ever heard of us!” Jacob joked. But Jacob and colleagues are on a mission “to end diagnostic odysseys.” His group has switched to an Illumina HiSeq 2000 and moved to whole-genome sequencing, even though that “just created a whole series of new problems” in data interpretation.
The move to push genome sequencing into routine clinical practice requires as much skill in navigating bureaucracy as interpreting base pairs. “Trying to get this CAP- and CLIA-certified is a nightmare,” said Jacob. The Sanger sequencing validation alone required 547 pages of documentation. But there’s no choice: “If you want your clinical colleagues to use this, they want this certified,” said Jacob.
Jacob’s team has jerry-rigged a series of data analysis guidelines, dubbed Illumina Carpe Novo (“seize the new”), designed for clinical geneticists to analyze whole-genome sequences, not as a research tool.
Since the Volker case, Jacob’s team has taken on six more cases, five of which had been sequenced and were undergoing analysis as of early March. Cases, which must be deemed actionable, must be nominated by two physicians. Following case review and consent (which takes 6-8 hours on average), the patient’s genome is sequenced, the data analyzed, and counseling offered.
The sequencing step must be clinically useful, said Jacob. “This is not for research. Case selection is based on the interest to the patient, not scientific knowledge.” Another tenet is equal access: cases are selected regardless of the family’s ability to pay. Jacob is focusing on putative monogenic disorders that give rise to a distinctive phenotype. “We’re not ready for common complex diseases,” he said. Moreover, “WGS must have the ability to enhance medical decision making, not make a diagnosis. We don’t make a decision [solely] on a variant in the sequence,” he said.
The review committee, which meets monthly, is chaired by the hospital’s chief medical officer, and includes ethicists, a genomics expert, and three external physicians. Some cases might be approved, others held in a “parking lot,” while others require further testing. Even after the Volker episode, Jacob was told that physicians were too busy or uncomfortable to recommend genome sequencing. But since October 2010, the committee has been reviewing 5-6 cases per month.
Another thorny issue is data return and retention. What information is given back to the patient or family? The ethics team decided that data return was “morally permissible” and “should remain at the discretion of the informed parental choice.” Parents are asked from the beginning which data they want returned.
As for the conviction that no insurance company would pay for genome sequencing, Jacob said he’d placed a $100 wager with a British colleague that two insurance companies would cover the costs by the end of 2011. “He owes me $100!” said Jacob.
Even more remarkably, Jacob has been approached by an unnamed insurance company interested in collaborating to develop “an evaluation clinic” to assess clients before expensive diagnostics testing. The company wrote: “In situations where you determine that on average, the cost of routine testing will exceed the current contract price of whole-genome sequencing, we will authorize whole-genome sequencing as a first-line clinical test . . . We are excited about the ongoing clinical utility monitoring you have established as part of your clinical whole-genome sequencing program.”
Whole-genome sequencing might sound like overkill, but Jacob argues it can be highly economical for two reasons. First, obtaining a complete genome sequence can cost less than running a handful of single-gene tests (some of which costs thousands of dollars). “I think we’ve overlapped,” said Jacob. “I’m going for the once versus every single gene. It’s simply easier to do the whole [genome].” And of course, the pay-off should a causative, actionable mutation be discovered could save (tens of) thousands of dollars in unnecessary medical procedures.
Jacob says he’s not scared by the falling cost of data generation. Data management and education around clinical decisions, on the other hand: “that gives me heart palpitations.” The security of the data remains an issue. Patient families are naturally concerned about who will see the data and will the data be in the public record? Physicians don’t want to be liable for interpreting the rest of the genome. Jacob’s team has concluded that the genome sequence and variants are not part of the electronic medical record.
Nicholas Volker didn’t eat solid food for nine months, but before he left hospital, he was eating steak with A1 sauce. He is now home and “doing quite well.”
Whole-genome sequencing is just “another lab value,” said Jacob. “It’s not a crystal ball until we prove it’s a crystal ball. That doesn’t mean it doesn’t have utility. Let’s not forget: individualized medicine is about choices. We think data is part of that.”
“Whole genome sequencing has the ability to effect change,” said Jacob. As for the full cost of Volker’s analysis, he said: “I have no idea—and I’m not going to count.”