By Monica Heger
This story was originally published on April 25.
A pair of studies published last week in the Journal of the American Medical Association highlight a host of regulatory and ethical issues confronting whole-genome sequencing as it moves into the clinic.
The two studies — one in which researchers used real-time whole-genome sequencing of a patient to determine the next course of treatment, and another that used sequencing to uncover an unexpected mutation with important health implications for the patient's children — illustrate the range of complex questions that will likely arise as the technology is adopted more broadly in clinical practice.
Both of the studies were done by researchers at Washington University in St. Louis on patients with acute myeloid leukemia. In the first case, a 39-year old woman with AML was demonstrating conflicting symptoms — some of which suggested an unfavorable outcome and that she needed a bone marrow transplant. Whole-genome sequencing was able to determine that she did not in fact need a bone marrow transplant, and suggested an alternative drug therapy that has since been used successfully.
In the second case, as part of a larger study on AML, the researchers sequenced the tumor of a woman, now deceased, who had developed aggressive breast and ovarian cancer at a young age, and then eventually developed AML due to chemotherapy. The sequencing revealed that the woman had a de novo germline mutation associated with a high risk of early-onset breast and ovarian cancer, the discovery of which has immediate health implications for her children.
The projects revealed that not only can whole-genome sequencing be used in a clinical setting to determine treatment, but also highlighted the importance of having in a place both an ethical and regulatory framework for conducting these types of studies.
"We and others are getting close to thinking about doing this work in real time, but [the regulatory issues] haven't been solved yet," Timothy Ley, associate director of Wash U's Genome Institute and a senior author of the paper, told Clinical Sequencing News. "It needs to be solved quickly."
Boris Pasche, a physician and professor of hematology and oncology at the University of Alabama, Birmingham, who wrote an accompanying editorial to the two JAMA studies, added that "the time has arrived" for clinical sequencing. "We are there.… That's the major message, and the very exciting finding" of these studies.
The 39-year old woman with a difficult diagnostic case of AML was referred to Wash U by her physician for consideration for a bone marrow transplant. The woman's case was difficult to diagnose because her cytogenetic profile suggested she had a complex pattern of chromosomal rearrangements, which is associated with poor survival rate, and typically treated with stem cell transplantation during remission. However, fluorescence in situ hybridization indicated a possible gene fusion, although not the one typically associated with AML. She was referred to Wash U during her first remission and the team performed another FISH test, as well as RT-PCR, to detect the gene fusion and did not find any evidence of it.
Her physician asked for whole-genome sequencing to be performed to see if it could identify clinically relevant information that would help determine whether she should receive a bone marrow transplant.
Ley said this case was the first time the team had performed sequencing and analysis of a patient in real time. "We usually don't operate under a clock when we're sequencing and analyzing genomes," he said. "We had no more than a couple of months to get the information needed to make the right decision."
Sequencing detected a rare oncogenic fusion that was undetectable using RT-PCR or FISH. The fusion involved the RARA gene, involved in other fusion events in AML, and for which drugs exist that are successful at treating those cancer subtypes. So while the initial clinical evidence suggested the patient had a poor prognosis and required a stem cell transplant, whole-genome sequencing revealed that the patient could actually be treated with available drugs and had a much better prognosis, Ley said.
Since the sequencing, the patient has started the therapy and is responding well.
The sequencing also enabled the team to create a new diagnostic for AML. "It gives us a scientific basis for how this [fusion] occurs and suggests that we need to look for it in other patients," Ley said.
Because the fusion was so small, commercially available probes used in diagnostics were not able to detect it, but Ley and his team have since created new probes that can detect the fusion.
The study also highlights the need for regulatory guidance. The Wash U team has in place a "moveable firewall," which maintains patient privacy, but if a medically relevant finding is found, that result can be returned to the patient's physician.
The team also validates all of its findings by creating a custom array for target capture and doing deep digital sequencing either on Roche's 454 platform or the Illumina platform. However, in order to return results to a patient they have to be validated on a CLIA-certified test, and the sequencing at Wash U is not done in a CLIA-certified lab.
In this case, the patient's RNA was severely degraded, making a confirmatory test on RT-PCR difficult, although ultimately successful, said Ley.
"We were lucky this time, but going forward we have to have a different tack," Ley said. Had the team not been able to confirm the results in a CLIA-certified manner, it is unclear whether they would have been allowed to return the results to the patient.
For the specific patient, "it's a moot question," he said, but, "can we do this routinely? No. Not until we get the sequencing into a certified setting."
[ pagebreak ]
In the second study, the researchers sequenced the tumor of a patient who had developed AML as a result of chemotherapy. When she was 37 years old, she had developed breast cancer, and at 39 was diagnosed with ovarian cancer. Despite being treated with surgery and chemotherapy, the cancer returned at age 42. She was treated with chemotherapy again, and sixth months later presented with AML. Because of the early age of onset, her physicians suspected that she had inherited a cancer susceptibility gene, but she had no family history of cancer and testing the most common cancer susceptibility genes yielded negative results.
When the Wash U team sequenced her genome, they discovered a deletion on TP53, a known tumor-suppressor gene. The deletion was homozygous in the leukemic DNA, and heterozygous in the patient's matched normal DNA. They then examined blood from the patient's mother and found only wild-type copies of TP53, determining that the patient's mutation was de novo.
Because the mutation is associated with a high risk of early onset cancer, the finding has immediate implications for the patient's children, who may have inherited the mutated gene.
"We were not expecting to find a familial predisposition gene for cancer," said Ley. But, whole-genome sequencing "is agnostic. So whether you're thinking about it or not, you get the result."
Again, because of the "moveable firewall" the team has in place, they were able to inform the patient's physician, who then contacted the family. A genetic counselor is also involved in explaining results to the patients as part of Wash U's protocol, said Ley.
To date, most of the cancer sequencing that has been done at Wash U has not been for the purpose of informing patients, because in many cases the patient is no longer alive. The team has been banking AML samples for the past 10 years.
"So the purpose of the studies has not been to return results to patients in real time," Ley said. "Going forward, we'll have to give a tremendous amount of thought to how we will return information."
Ley said that he and his colleagues classify mutations into three different categories. The first is a mutation that will have an immediate impact on the patient and patient care, such as the mutation that was found in the first case that determined that the AML patient did not need a stem cell transplant. The next is any mutation that is known to increase the risk of cancer. And the third level is any "incidental finding that is completely unexpected, [such as] variants relevant for other diseases."
A mutation that falls in the first level would be returned automatically, said Ley, while the next two levels of mutations should be up to the patient.
"Our general approach for incidental findings is to ask the patient whether they want the findings returned. We make people aware that this could happen during whole-genome sequencing. … We ask the patient whether they want that info to be shared, and if they do, we share it, with a genetic counselor guiding the process."
He said that the moveable firewall protocol that the team had set up about six or seven years ago worked well in both of the cases. "It worked as we had intended it to work," he said. "We didn't have to invent something new."
However, the studies did illustrate the need for a regulatory framework for doing whole-genome sequencing in a clinical setting. At the moment, it's not clear what such a framework would look like, he said.
Moving forward, Ley said reading through the noise, making sense of the data, and figuring out a path for delivering results to patients will be the main challenges in implementing clinical sequencing on a broad scale.
"One of the biggest challenges is to sift through the noise and find the [mutations] relevant for diagnostics, prognosis, and targets," he said.
Additionally, the cost is currently prohibitive in most cases; however for diseases like AML, where the cost of treatment can be enormously high, the technology is beginning to make sense. A stem cell transplant could have cost upwards of $500,000 to $1 million, he said, so spending $10,000 to $20,000 to sequence the genome is a comparatively small investment, and the potential for benefit is strong.
"The beauty of whole-genome sequencing is it's a one-size-fits-all platform. It's a single platform for testing all patients," he said, as opposed to devising 40 or 50 different diagnostic tests for different cancers.
Cost is the only barrier to clinical sequencing, according to Pasche. While some researchers and physicians have expressed concern over interpreting sequence results or incidental findings, Pasche said that at least in oncology, any technology that will enable more personalized and targeted therapies will be extremely welcome.
As an oncologist who treats patients with gastrointestinal malignancies, he is excited by the prospect of sequencing cancer genomes and how that will impact treatment, he said, adding that already, developing personalized drugs based on a person's genetic background has led to improved survival rates.
"We are obviously using [those genetic tests] already and there's been no resistance in the oncology community to using them," he said.
Have topics you'd like to see covered by Clinical Sequencing News? Contact the editor at mheger [at] genomeweb [.] com.