Arup Laboratories has joined the ranks of academic and commercial laboratories offering clinical exome sequencing to diagnose unknown diseases.
In a presentation at the Cambridge Healthtech Institute's Clinical Genome Conference in San Francisco last month, Elaine Lyon, medical director of the genetics division at Salt Lake City-based Arup Laboratories, said that the laboratory has decided to follow the American College of Medical Genetics and Genomics' guidelines on returning incidental findings. She made her comments in both a presentation at the conference and in a follow-up interview with Clinical Sequencing News.
The ACMG guidelines, which were released in March, recommend that laboratories doing clinical exome or whole-genome sequencing search for known pathogenic variants in 57 disease genes related to 24 disorders and return those results to the ordering physician regardless of patient preference.
While most labs have instead decided to give patients the option to not have those results returned to the physician (see CSN 5/8/2013 and CSN 5/29/2013), Arup is one lab that is returning them. In addition, Lyon said that it is considering adding some pharmacogenomic variants to the list, which are not included in the ACMG recommendations.
Arup deems their exome test a "symptom-guided analysis," because following the sequencing, analysis is prioritized based on the patient's clinical symptoms and phenotype, Lyon said.
For Arup's exome test, sequencing is done using Agilent's SureSelect and the Illumina HiSeq 2000. Turnaround time is about three months, and Lyon said that pricing strategies were proprietary.
Arup launched the test in May 2012 and requires a physician prescription, but other than that there are no restrictions on who can order the test or for what indication, Lyon said, as long as it is a suspected genetic disease. However, she noted that the laboratory vets each case to make sure it is appropriate for exome sequencing. Thus far, Lyon said the majority of patients have presented with a neurological condition or developmental delay.
Thus far, the laboratory has found the causative mutation in 27 percent of cases, and in another 27 percent has found variants that may be associated with the patient's condition, Lyon said.
Arup bills the hospital laboratories that order the test directly and does not deal with the insurance companies, so Lyon did not know how often or how much of the test was being reimbursed.
Laboratories such as Baylor, the Medical College of Wisconsin, and the Mayo Clinic have reported increasing success in obtaining reimbursement, although the amount that is reimbursed is variable and is on a case-by-case basis.
Key features of Arup's exome pipeline and what Lyon dubs the "symptom-guided analysis," are tools the laboratory has developed to analyze and classify variants — VarViewer and VarRanker.
VarViewer enables the researchers to set parameters around the types of variants that are analyzed. "It uses variant frequencies, family inheritance patterns, and mutation databases," Lyon said.
Then, VarRanker ranks those variants in terms of the ones that are likely to be causative for the patient's symptoms. "VarRanker scores the variant using key terms in from the patient's clinical findings, gene ontology, PubMed abstracts and the STRING protein interaction database," Lyon said. Additionally, it looks for variants that damage the gene and that are in a pathway related to the phenotype. "
Because of Arup's analysis approach in using the patient's phenotype to guide analysis of the exome, Lyon said that the lab may miss many incidental findings. Instead it does a separate analysis to identify incidental findings.
Prior to the ACMG recommendations, she said that the team at Arup had been discussing returning incidental findings and had been trying to come up with their own list of genes that should be returned to the physician in all cases. However, aside from the ACMG list, the laboratory returns only variants related to the patient's phenotype.
"We felt there is an obligation to return," she said, "but we were uncomfortable coming up with our own panel and appreciate the fact that we have some guidance from a professional society." She said the lab chose to follow the ACMG recommendations and return such findings to the physician in all cases because "the guidelines put it with the physician, who has more contact with the patient, to return those results," she said. "It's difficult for labs to manage that because we don't have the direct contact with the patients."
Moving forward, she said the lab is looking to see whether it can validate pharmacogenomic variants, and if so, will return those to the physician as well. However, many of those variants are difficult to analyze with next-generation sequencing, Lyon said.
For instance, the CYP2D6 gene, a common gene involved in drug metabolism, contains a number of variants that are difficult to detect with next-gen sequencing including deletions, duplications, and gene inversions. In addition, it also has a pseudogene, and "we would need to make sure we're not identifying variants in the pseudogene instead of the real gene," Lyon said.
Thus far, Lyon said that one of the main challenges of the test has been interpretation. There is a "need for caution" in interpretation, she said. False positives are especially problematic, she said, because patients will be given the wrong diagnosis, and family members will be tested and also be given wrong diagnoses.
At last month's Clinical Genome Conference, she presented on several tricky case examples. In one example, the patient had phenotypic symptoms of hereditary hemorrhagic telangiectasia, or HHT disease, an autosomal dominant disease that causes abnormal blood vessel formation. Symptoms of the disease can range from nosebleeds to more serious problems like vascular malformations in organs and intestinal polyps.
Exome sequencing identified a "great candidate" — a stop codon mutation that tracked with the disease in the immediate family. However, the laboratory also happened to have an extended family tree and on further analysis, "it was clear that the variant did not track with the family members," she said. So the variant was ruled out as being the cause. She said this example illustrates the fact that diagnostic exome sequencing "needs to be approached carefully and cautiously."
Another key piece is informed consent. Lyon said it is important for families to understand the full benefits and risks, including the possibility that the test will not yield an answer or could reveal unexpected family relationships. Often, laboratories have to "deal with the fact that the relationships are not as they were told to us," she said.
Lyon estimated that around 25 percent of patients receive a definitive diagnosis from an exome test, and that while this metric will improve as more patients are sequenced, it is important for families to understand that.
As clinical exome sequencing continues to become more widespread, interpretation will become easier, she said. Additionally, "the technology will continue to improve, so calls will be more solid and we'll be able to read through the more difficult portions of the genome."
Particularly for patients with rare diseases without a molecular diagnosis, exome sequencing is already proving to be cost-effective, she said. "If you look at the amount of testing these individuals go through in trying to come up with a diagnosis, an exome test is becoming fairly reasonable."