NEW YORK (GenomeWeb) − Genetic tests for hereditary cancer involving one or a few genes are starting to be replaced by multi-gene sequencing panels. But the clinical utility of the new tests, which have been enabled by advances in next-generation sequencing technology, has been unknown.
Now, recent studies in various patient groups using panels from Ambry Genetics, Myriad Genetics, InVitae, and an in-house test suggest that multi-gene panels increase the diagnostic yield by about 10 percent over more targeted tests and sometimes deliver unexpected findings in genes not previously associated with a certain cancer type. However, no clinical guidelines exist for many of the genes included in the panels, and the rate of variants of unknown significance is currently high.
Last month, Ambry Genetics published a study in Genetics in Medicine that reported results from the first 2,079 patients tested with its commercial BreastNext, OvaNext, ColoNext, or CancerNext hereditary cancer panels, which cover genes associated with breast, ovarian, and colon cancer as well as across cancers. The panels, which Ambry started offering in 2012, included between 14 and 22 cancer susceptibility genes but not BRCA1 and 2.
Each panel had a diagnostic yield of almost 10 percent, either a pathogenic or likely pathogenic mutation in one of the genes: 7.4 percent for BreastNext, 7.2 percent for OvaNext, 9.2 percent for ColoNext, and 9.6 percent for CancerNext.
According to Liz Chao, Ambry's chief medical officer, many of the results had clinical utility because they provided families with diagnoses, explained their cancer risk, and allowed additional family members to be tested to determine their risk. For the cancer-specific panels, she said, 90 percent of the mutations were found in genes that made sense for the phenotype seen in the patient.
One of the most important findings, she said, was that many patients with a pathogenic mutation in a particular gene would not have qualified for testing of that gene by phenotype-based diagnostic criteria, so without the panel, they would not have been diagnosed. "We're probably missing between a quarter and a third of patients when we use a gene-by-gene approach, so there is clear utility of the panels in terms of increasing the diagnostic yield," she said.
But many of the genes on the panels currently do not have clinical guidelines for managing cancer risk associated with them, such as prophylactic surgery or chemotherapy. For example, eight out of the 14 genes on the breast cancer panel don't have such guidelines. "It shows the state of where the research is today," Chao said. "We don’t have enough data, particularly prospective data, to develop these guidelines."
Ambry is working with academic groups, including the University of Pennsylvania and Memorial Sloan-Kettering Cancer Center, on a prospective study of patients with mutations in genes that do not currently have guidelines in order to define their disease penetrance. That study, she said, is expected to start recruiting patients over the next several months.
The panels also had a high rate of variants of unknown significance, ranging from 15 percent to 25 percent. However, those rates have already come down since Ambry started offering the panels because of results from family studies and other research. "The more testing we do, the more the rates come down," Chao said. She also noted that VUS rates are still higher in non-Caucasian populations, which will change when more reference genomes from other ethnicities become available.
In the meantime, she said, it will be important that genetic counselors or clinical geneticists who are used to dealing with VUS are involved in conveying the results to doctors and their patients, and to encourage patients with VUS to get enrolled in family studies that can help to classify their variants.
Also last month, researchers at Stanford University, in collaboration with InVitae, published a study in the Journal of Clinical Oncology in which they tested the performance of a 42-gene panel from the company for assessing cancer risk in 141 retrospective samples from women with a family history of breast or ovarian cancer who had previously tested negative for mutations in BRCA1 and BRCA2.
The study found pathogenic variants in about 10 percent of the women who received their results. Most of those findings prompted a change in their care, in particular enhanced screening. In one case, that led to the early detection of a precancerous colon polyp.
Notably, not all of the mutations were in genes associated with breast or ovarian cancer – they also included genes associated with colon and gastric cancer. "If you find one of these genes in a family that doesn't really fit the typical criteria for what that gene usually causes, how do you interpret that?" said Jim Ford, a professor of medicine and genetics in the division of oncology at Stanford University School of Medicine, and the senior author of the study. "Is it not as risky? Just a weird chance? Is it not actually pathogenic? We really don't know the answer to that."
Because of this, not all likely pathogenic mutations were clinically actionable. For example, in one patient, the researchers found a mutation in the gene CDH1, which has been associated with hereditary gastric cancer, and the usual recommendation is a prophylactic gastrectomy. However, because the mutation was not found in a family with a history of gastric cancer, the researchers did give that advice. "That just seemed way beyond what was reasonable," Ford said.
Similar to Ambry's panels, some genes on InVitae's panel have no clinical guidelines associated with them, and for many genes, there is no good data on their frequency in different populations and their penetrance.
To get a better grip on this, more studies are needed, in particular population-based studies that include not only patients at high risk for a certain type of cancer but patients across different cancers, Ford said.
Stanford, the University of Southern California Norris Comprehensive Cancer Center, Myriad Genetics, and others are planning a large multi-year prospective clinical study of Myriad's myRisk cancer panel and its clinical utility, he said. The study was recently approved and is expected to open this summer.
In addition, to learn about the impact of mutations in genes with moderate penetrance, studies of tumors need to be conducted, Ford said, in order to prove that an inherited mutation was actually involved in a tumor's formation.
Like Ambry's study, the Stanford and InVitae researchers found a high rate of variants of uncertain significance – an average of 2.1 per participant. While that rate was not a surprise, given the large number of genes, "it is difficult because we don't know what to do with those," Ford said. VUS don't have any clinical impact – they are treated as negative results – "but they're very confusing for patients and health professionals alike."
With more patients tested, and more data across ethnic groups, VUS rates are expected to come down quickly over time. A follow-up study with InVitae, for example, that involved a similar number of patients and a smaller panel of 29 genes, already found a substantially lower VUS rate, Ford said.
Data sharing among laboratories could also help bring the VUS rate down. "From a scientific point of view, the more VUS [that] are at least in the scientific domain, the faster we will be able to interpret these things," Ford said.
For now, while multi-gene panels have a place in the clinic and are already being offered commercially and ordered by physicians, they should only be used in select situations, Ford said. "I think they should be used with caution and in a setting with expertise in genetic counseling and cancer genetics at this point as we learn more about them," he said.
"I don't think it's something you just order on anybody who has a cancer," Ford said. They're really for high-risk families where the standard testing has not yielded something. In that setting, about 10 percent of the time, you can find something else."
In the meantime, Myriad Genetics plans to phase out its single-syndrome tests by the summer of 2015 and replace them with its 25-gene myRisk hereditary cancer panel. MyRisk has been available to clinicians under early access since last September and will be generally available by the end of this summer.
Several researchers will present their initial experience with Myriad's myRisk at the American Society of Clinical Oncology annual meeting in Chicago, which starts later this week.
Researchers at the Dana-Farber Cancer Institute in collaboration with Myriad, for example, tested myRisk in a retrospective study of 1,260 patients with a history of Lynch syndrome who had previously undergone testing for five Lynch syndrome-specific genes. According to the meeting abstract, the myRisk test found pathogenic mutations in 160 of those patients.
Seventy-three patients had mutations in Lynch syndrome genes, and 30 percent had mutations in one of the 20 other genes on the panel. Of those, almost a third had mutations in the BRCA1 or BRCA2 genes, an unexpected finding because none of those patients had breast cancer.
Those findings, though they were not communicated to the study participants because of the study's rules, could help prevent cancer in additional patients. The results are "going to accelerate the switch from single-syndrome testing to myRisk testing," and will likely be "practice-changing," said Richard Wenstrup, chief medical officer of Myriad Genetics.
Like in the published studies, the VUS rate was high in this one – 44 percent of patients had at least one VUS – but that was not unexpected. BRCA1 and BRCA2 testing also started out with VUS rates of 30 percent, now down to 2 to 3 percent, Wenstrup said. "It's completely unavoidable, but we feel we have a way forward to reclassify them and to reduce this rate," he said.
Myriad expects hereditary cancer panel testing to become widely accepted in the near future. "When people start to see really important mutations that they would be missing if they were using a single-syndrome approach, we think that the field will be changing fairly quickly," Wenstrup said.
Another study to be presented at ASCO comes from the University of Pennsylvania. The retrospective study involved 278 patients with early-onset breast cancer that had previously tested negative for mutations in BRCA1 and BRCA2 and used a 22-gene panel developed by the researchers.
Eleven percent of patients were found to have at least one deleterious or likely deleterious mutation in one of the genes, but only 2.5 percent of patients had such mutations in genes for which clinical guidelines for management exist. Like the other studies, the VUS rate was high, 19 percent for genes with high or moderate penetrance.
More studies are needed for the genes without clinical guidelines, according to Kara Maxwell, a fellow in the division of hematology-oncology at the University of Pennsylvania School of Medicine, who led the study. "That is really, in our opinion, the major focus of what needs to happen next, to help establish what we should do with these various genes. We have to define penetrance, and frequency of mutations, and a lot more work needs to be done," she said.
And for those genes that have well-defined guidelines for one type of cancer, their involvement in other types of cancer needs to be studied. "What do you do when you find a mutation in a family that is not classically associated with mutations in that gene, how should you apply the existing guidelines in that scenario?" she said.
Like Stanford's Ford, Maxwell is cautious about the clinical use of hereditary panel testing, though it is warranted in certain cases. "There is such a benefit to potentially understanding the genetic risk in your family that for individuals that have a significant family history and are considering this kind of testing, I'm certainly not saying not to do it," she said. "The recommendation I would not make is to consider this kind of testing outside of a large academic institution with a genetics team that can help interpret it."