SAN DIEGO (GenomeWeb) – Following several years of case examples where exome or whole-genome sequencing has led to a diagnosis and sometimes even pointed to a life-saving treatment for a child with a rare, debilitating genetic disorder, evidence is now building demonstrating that the technology can be a powerful tool when prescribed as a clinical test for patients with a genetic disease with an unknown cause.
Two recent studies published in the Journal of the American Medical Association described the results from exome sequencing of over 2,800 patients in clinical laboratories at Baylor College of Medicine and the University of California, Los Angeles. The studies corresponded with the annual American Society of Human Genetics meeting, held here this week, where additional laboratories reported on their experiences with implementing clinical exome sequencing.
The Baylor and UCLA teams reported in the JAMA studies that on average exome sequencing led to a diagnosis in about 25 percent of the patients, although certain clinical phenotypes were more likely to yield a diagnosis than others. In addition, including exome data from both parents boosted the diagnostic rate to 31 percent. Actionable incidental findings were found in 3 percent to 5 percent of patients, depending on the criteria for defining an actionable gene. And overwhelmingly, patients wanted to know the results of those incidental findings.
Exome sequencing is "proving to be quite useful for the molecular diagnosis of very difficult to diagnosis kids," Stan Nelson, a professor of human genetics at UCLA, told Clinical Sequencing News in a phone interview. He thinks that studies such as these will help build evidence for the technique and eventually lead to exome sequencing being done earlier in the quest for a diagnosis. "I think it will guide [physicians] to think about molecular testing perhaps in a different order. Many of these patients came to us after thorough genetic testing," Nelson said. Going forward, "the flow is going to be that exome sequencing will be introduced much earlier in the process."
During a presentation at the ASHG conference, Christine Eng, medical director of the DNA Diagnostic Laboratory at Baylor, said that there is "strong, growing interest" for the exome sequencing test and that the lab has now sequenced 5,100 samples and receives around 230 samples per month.
"We diagnose rare conditions as well as some common ones," she said.
In addition, she said that of the first cohort of 250 patients that the lab previously published on last year in the New England Journal of Medicine, new gene discoveries in the subsequent years have increased the diagnostic rate of that group from 25 percent to 28 percent.
"New gene discoveries are definitely going to drive our diagnostic rate higher," she said. Improving exome coverage will also lead to more solved cases, she said, but there are also likely variants in non-coding regions of the genome that will not be picked up by exome sequencing.
Of the 2,000 cases referred to Baylor for exome sequencing, 900 were under the age of five, 845 patients were between 5 years and 18 years, 244 were older than 18 years, and there were also 11 fetal samples from terminated pregnancies. The vast majority, 87.8 percent, had some type of neurological disorder, while only 12.2 percent had a non-neurological disorder.
A total of 708 causative variants were identified in 504 cases, more than half of which were novel, for a diagnostic rate of 25.2 percent. In addition, 30 percent of patients had mutations in disease genes that had only been recently described in the literature.
The UCLA team took a slightly different approach, preferring to sequence the exome not only of the patient, but also of both parents. From 814 cases referred to the center for testing between January 2012 and August 2014, 213 cases, or 26 percent, received a diagnosis.
Trio sequencing increased the diagnostic rate significantly, with 127 out of 410, or 31 percent, of those cases receiving a diagnosis. Of the 338 non-trio cases, only 22 percent were diagnosed.
The difference between trio and patient-only sequencing was even more pronounced for patients presenting with developmental delay. For those patients, trio sequencing had a 41 percent success rate but there was only a 9 percent success rate when parental samples were not available. The authors attributed this difference to the ability of trio sequencing to identify de novo and compound heterozygous variants.
"Our strong preference is for physicians to send trios," Nelson said. Trio sequencing facilitates analysis and is more likely to lead to a diagnosis, although sequencing the exomes of the parents also has its own issues, including cost and the potential of incidental findings or uncovering unexpected familial relationships, he added.
Nelson said that sequencing trios is about 50 percent more expensive than sequencing just the patient because the majority of the cost is the analysis, not the sequencing, and trio sequencing reduces the bioinformatics burden.
An ongoing issue with clinical exome sequencing is the possibility of finding mutations in disease genes not related to the patient's current condition. In the Baylor study, the group found such results in 4.6 percent of patients. When considering the 56 genes that the American College of Medical Genetics and Genomics recommends reporting on, about 3 percent of patients had actionable variants. An additional 33 patients, or 1.7 percent, had mutations in other genes that meet Baylor's requirements for medically actionable.
If medically actionable results are identified in the patient, the parents have the option of ordering testing for themselves free of charge, and to date, 33 parents from 19 families of those 92 patients have requested testing.
Baylor uses a two-tier strategy for reporting its clinical exomes. Tier 1, which is the more focused report, includes six categories: pathogenic variants related to the disease phenotype; variants of unknown significance related to the disease phenotype; medically actionable mutations in genes not related to the disease phenotype; carrier status for autosomal recessive genes recommended for population-based screening; a limited number of pharmacogenetic variants; and clinically relevant mitochondrial variants. Tier 2 includes deleterious mutations or variants of unknown significance in genes not related to the disease phenotype and predicted deleterious mutations in non-disease genes.
As of September 2013, Baylor gave patients the option to decline results in the Tier 1 report for non-ACMG incidental finding genes.
Overall, 1,808 families requested all aspects of the focused report. Only 128 and 154 declined to receive information on recessive disorders and pharmacogenetic variants, respectively. Of the 190 families given the opt-out for non-ACMG incidental findings, only one opted out. In addition, the physicians of 524 patients requested the larger Tier 2 report.
UCLA has slightly different criteria for deciding what genes to include in its analysis of incidental findings. The variant must be in a gene unrelated to the patient's phenotype and also predicted to be pathogenic according to current ACMG standards. While patients were allowed to opt out of all incidental findings, 97 percent of families have chosen to receive them. In total, the UCLA team found incidental findings in 260, or 3.2 percent, of cases.
Nelson said that he was somewhat surprised by the fact that so many parents wanted to receive incidental findings. The UCLA group was opposed to ACMG's initial recommendation that laboratories offering clinical exome sequencing should search for and report back on pathogenic variants found in a list of 56 disease-related genes, regardless of patients' preferences for having those results. "I and our team believed in giving patients autonomy," Nelson said. The ACMG has since revised its guidance to support patients' choice of whether or not to receive those results.
Nelson said that when UCLA first began offering clinical exome testing and wanted to do trio sequencing, many parents commented that they would only consent to having their exomes sequenced if they did not have to know their own results. They just wanted the child's diagnosis. "So, I was assuming that there would be a larger fraction of people that wouldn't want their incidental findings," Nelson said. "But it turns out that very few people opt out."
Labs report comparable results
Baylor and UCLA are not the only labs to offer clinical exome testing, and at the ASHG conference this week, several commercial laboratories reported on their experiences with the test.
At a workshop sponsored by Ambry Genetics, Layla Shahmirzadi, a genetic counselor with the company, said that of its first 500 patients submitted for exome sequencing the company reported back a molecular diagnosis for 37 percent of patients. For 7 percent of those patients, the diagnosis was reported in a novel gene, she said. Ambry offers two different exome tests, both of which are trio sequencing. The standard exome test costs $5,800 and has a turnaround time of eight to 12 weeks, while its rapid exome test costs $15,129 and has a turnaround time of two to five weeks.
At the University of North Carolina, Chapel Hill, a project called NCGENES —sponsored by the National Institutes of Health under its Clinical Sequencing Exploratory Research arm — is sequencing the exomes in patients with a range of conditions to evaluate exome sequencing as a diagnostic tool.
Natasha Strande, a postdoctoral research associate in Jonathan Berg's lab at UNC, reported that of the first 300 patients the group has sequenced, it has made a positive diagnosis in 17 percent of cases and a possible diagnosis in an additional 23 percent of patients. Patients have a range of disorders that are of suspected genetic origin, including hereditary cancer, cardiovascular disorders, neurological disorders, retinal disease, and dysmorphology.
Similar to the Baylor and UCLA group, the UNC team has found that the diagnostic yield varies based on phenotype. Patients with suspected hereditary cancer have the lowest diagnostic rate, at around 13 percent, while ophthalmology- and cardiology-related disorders are each diagnosed more than 50 percent of the time, Strande said.