NEW YORK (GenomeWeb) — About two years from its launch in 2012, a pediatric cancer sequencing study led by Baylor College of Medicine and Texas Children's Hospital has recruited more than half of its intended 280 subjects and is beginning to collect data on the effectiveness of exome sequencing in finding actionable mutations that might inform more personalized treatment for these patients.
Will Parsons, a Texas Children's pediatric oncologist who co-leads the project with Baylor geneticist Sharon Plon, discussed the study's progress in a session at Cambridge Healthtech Institute's Next Generation Diagnostics Summit, held last week in Washington, DC.
The Baylor and Texas Children's study was one of several to receive funding under the National Human Genome Research Institute's Clinical Sequencing Exploratory Research program.
As part of their CSER project, the Baylor and Texas Children's researchers are evaluating not only the utility of whole-exome sequencing in guiding the care of children with cancer, but also tracking the process on the ethical, sociological, and practical side. To this end, not only children, but also their parents and their oncologists are enrolled as subjects in the study.
Parsons said in his talk that the team has recruited about 165 newly diagnosed pediatric cancer patients, more than half of its intended 280 subjects. Along with these kids, about 220 parents have also enrolled, as well as the study's 16 participating oncologists.
The researchers are studying the diagnostic yield of exome sequencing, and how often it impacts the decision-making of participating oncologists. They are also studying the process and impact of consenting families to have their children's exomes sequenced, and the process of returning both cancer-specific results and germline findings potentially incidental to a child's cancer but actionable in some other way.
By analyzing counseling session transcripts, the team hopes to inform best practices by studying how patients and their families respond to the process.
Parsons told Clinical Sequencing News this week that so far, enthusiasm from families has been robust. A majority of those approached, about 83 percent, have consented to join the study. Only 17 percent have declined.
Parsons said at the meeting that families' interest has been so robust that a significant part of the study consent process has actually been "dampening that enthusiasm." The study's counselors have to explain to families "Yes we can identify potentially useful results, but at this point we don’t know that we will, or that they will actually be used," he said.
The group is also fielding interest from families in exome sequencing outside of the study protocol, Parsons told CSN, but currently it only rarely orders exomes in the clinical setting.
Within the study, researchers are already getting hints at how often exome sequencing is able to find a potentially actionable mutation in children already enrolled.
In about 85 to 90 percent of their solid tumor cases, and 70 percent of their CNS cases, the researchers have been able to obtain tumor tissue to sequence, Parsons said in his presentation.
In its reporting, the study team decided to divide tumor variants into four categories: the first making up mutations that appear to be clearly important or actionable in the context of the patient's particular cancer; the second covering alterations that have less direct or more uncertain evidence for their utility; the third, mutations in consensus cancer genes that may not have any therapeutic relevance for a child's particular cancer; and the fourth, mutations with really no indication of usefulness.
As part of the study protocol, the group is also sequencing patients' germline DNA and reporting actionable results, except for recessive carrier status, which families can opt in or out of receiving. According to Parsons, virtually all have opted in.
In its early subjects, the study has seen exome sequencing identify some actionable mutation — either a category-one directly actionable cancer mutation, a category-two potentially actionable cancer mutation, or something else of clinical import in the patient's germline DNA — about 30 to 40 percent of the time.
According to Parsons, many of the more well-known actionable mutations have been present in a very small fraction of cases in the first subjects sequenced. The number of mutations found for each patient has varied widely, from around 80, down to very, very few.
Only a small percentage of subjects have had a category-one mutation. Another quarter of patients have had a category-two mutation. About another 23 percent of subjects have had a category-three mutation, and half of participants have had nothing there that could be interpreted as actionable.
Virtually every patient, Parsons said, has had some variants of unknown significance — two to three per subject, on average — as well as an average of two recessive carrier mutations and two pharmacogenomic variants.
The lingering question, Parsons said, is what all of this means in terms of optimizing clinical practice moving forward.
"We are seeing a diversity of mutations in a host of different genes, many captured on adult panels, but what's going to be the test of choice, a smaller panel or something more genome-scale? … Or, by adding other types of testing how can you improve upon this and determine what is the test you should actually use?" he said.
At the meeting, Parsons highlighted a handful of case studies from the project so far illustrating the somewhat mixed blessings of clinical exome sequencing, as well as hints the researchers have had of the importance of adjunct tools like RNA sequencing to help better understand exome sequencing results, and, in some cases, provide independent diagnostic information.
When possible, the researchers in the Baylor study are performing additional analyses like panel-based sequencing, RNA sequencing, and, in some cases, whole-genome sequencing to try to understand what the benefits and limitations of each may be.
In one of the cases Parson discussed, sequencing revealed a TSC2 mutation in a 14-year-old girl with poor-prognosis osteosarcoma. "This is obviously important," Parsons said in his talk. But it's not necessarily obvious what to do with it.
"These are the decisions that are going to be facing oncologists. Taking a rational approach with a patient with poor prognosis, how do you decide, if you don’t have a clinical trial of mTOR inhibition in osteosarcoma patients, what to do?" he asked.
In another case of a teenage boy with renal cell carcinoma, exome sequencing did not reveal any mutations that might inform therapeutic decision making, but when the researchers added RNA-seq, they found a previously unidentified ALK fusion.
"It's still a question that remains to be answered, what the optimal test is going to be realistically," Parsons told CSN.
Given that US Food and Drug Administration-approved drugs largely target genes in these common mutations panels, "you can make a case for that as a way to go," he said. "I think in the end though that, assuming cost continues to go down, something [broader] makes sense. We've already seen enough cases, for example, [with] RNA-seq where it identifies clinically relevant fusions that hadn't been seen," Parsons added. "[RNA-seq] is also really helpful to interpret DNA results, to see whether mutations are expressed or whether a particular pathway is activated."
In his talk Parsons also highlighted the case of a 14-year-old girl with glioblastoma and no recorded family history of cancer. After she was sequenced, however, the researchers discovered a germline MSH2 mutation, inherited from her mother. MSH2 mutations are some of the most common drivers of Lynch syndrome, an inherited cancer susceptibility disorder, and are associated with increased risk of gliomas.
"Obviously this had practical clinical importance for testing of other family members, as well as surveillance for the patient herself," Parsons said.
The group has also identified mutations explaining non-cancer health issues in study participants, including a pathological mutation linked to a child's unexplained liver disease, and another to a child's renal disease.
Over the next few years, as the study recruits its full cohort and continues to track the diagnostic yield of exome sequencing, as well as other approaches, Parsons said the group hopes to have a more complete picture that can help inform best practices in clinical sequencing of pediatric cancer patients.
The study is also going to look at how sequencing results impact oncologists' treatment decisions. Because first-line therapies for pediatric cancers are well established, the study is not intended to inform initial treatment. But the researchers will look at how sequencing might change treatment plans when patients relapse down the line.
So far, after just a few years, there is not enough data on this end to draw any conclusions, Parsons said. The team has tried to tease some of this out by surveying the participating oncologists before and after receiving sequencing results on whether the results change their potential treatment plans, but there's no sense yet how this will actually play out.
One insight that has already become apparent is that the resources and infrastructure of large institutions like Baylor and Texas Children's have been invaluable to sequencing the exomes of these pediatric patients, Parsons said.
"Because of the grant and the infrastructure we have, we are really working in a best case scenario," he said. "We have two wonderful genetic counselors funded by the study, one of whom works in our clinical cancer genetics clinic, so if there are patients that should be referred there, we can do that easily."
"We are also in this huge hospital so if there are incidental findings related to mitochondrial disease or arrhythmia we can refer patients directly there to clinicians in those specialties," Parsons said. "It's a very different scenario than an oncologist in a small practice outside an academic medical center."
The group's experiences are also going to be an important background for moving toward sequencing pediatric cancer patients in the context of a prospective trial, where patients matched to therapy based on their genome are compared to those given standard therapy, something Parsons said the Baylor and Texas Children's researchers are planning to pursue.
"The idea is to use these technologies as part of a prospective trial akin to the [National Cancer Institute] adult studies like MATCH or M-PACT," he said. "We are currently working on how to do the same in kids, and a lot of this current study has been prepping for that more interventionist next step."