NEW YORK – Children's Mercy Kansas City has created a new laboratory-developed test based on long-read sequencing, which it says could reduce the number of tests needed to diagnose critically ill newborns.
The recently validated test uses so-called "five-base" HiFi sequencing from Pacific Biosciences which provides methylation calls in addition to long reads. Methylation assays for conditions such as Angelman syndrome are just one of the genetic tests that hospital officials hope to replace with long-read WGS, as well as karyotyping, clinical microarrays, and PCR for repeat expansions.
"Every additional test ranges from several hundred to thousands of dollars. We estimated that we can probably save, in this patient population, about one-third of testing costs by consolidating," said Carol Saunders, director of Clinical Genetics and Genomics at Children's Mercy, who led validation of the test. She noted that this calculation assumes these other tests would be performed for a patient, either in parallel to genome or exome sequencing or in series.
"But the key benefit is the fact that the babies and children get uniformly tested through all modalities and not selectively, based on physician intuition," she said. "As we go along with this HiFi-first for critically ill children, there may be large numbers of certain types of disorders that may have been missed before."
Now, Children's Mercy must study the frontline clinical implementation of the test to evaluate its utility in critically ill newborns. "We have the data to justify HiFi-first," Saunders said.
The hospital looks to analyze hundreds of patients — and their parents — to see what benefits the test may provide. Diagnostic yield could increase by 5 percent or more, compared to clinical exome sequencing, according to Tomi Pastinen, director of the Genomic Medicine Center at Children's Mercy. That's straightforward to measure, while other aspects of clinical utility may be exposed for the first time, such as time to diagnosis. "What are the potential things one may avoid by doing a single test?" he said. "I think those are what we want to learn. It's difficult to do in any context other than clinically applying it to successive patients."
Offering long-read sequencing as a first-tier diagnostic technology is "exciting," said Michael Bamshad, a researcher at the University of Washington and principal investigator of the SeqFirst study. His lab has partnered with PacBio and GeneDx to evaluate HiFi sequencing as part of a "bake-off" between long- and short-read WGS technologies. "It'll be very interesting to see what their experience is a year from now," he said.
Children's Mercy's LDT is the next step in its exploration of whole-genome sequencing for diagnosing sick children. Along with Rady Children's Hospital, it has pioneered the use of sequencing, and especially long reads. The decision to create the assay and study its use as a frontline test is based on years of research in the Genomic Answers for Kids program.
"The scientific part is done," Pastinen said. "Research data from over 1,400 cases we've sequenced here tell us that HiFi genome sequencing will find everything that short reads and other tests will find."
Bamshad, who is evaluating PacBio's HiFi sequencing as well as Oxford Nanopore Technology's platform in his study, suggested that long reads may have more noise when it comes to single-nucleotide variants. "The concern is there might be SNVs called by short reads, but missed by long reads," he said.
Perhaps the biggest challenge will be making the economic case for long reads. "It would have to demonstrate superiority to the extent it could justify increased cost," Bamshad said.
PacBio's new Revio instrument, which increases throughput and theoretically reduces the cost per 30X human genome to less than $1,000, has been key to developing the LDT, Saunders said. "In the Sequel II era it would have been infeasible," she said, referring to Revio's predecessor. "I don't think this would have been possible, even if we already knew there was scientific justification for some indications."
The validated test is for a trio — child and parents — and with the Revio reagent chips, all three genomes at 30X coverage can be done in one run. "It makes it easier to run continuous operations on demand," she said. "You don't need to accumulate samples for a run."
Reducing turnaround time to about two weeks was a goal of the LDT development process. "We call this expedited testing," Saunders said, "but we're not calling it rapid." Once the samples have been obtained, library prep takes around three days, with another one and a half days for the sequencing run. Primary analysis and interpretation take about two days combined. Then the results are turned over to clinicians to interpret and deliver findings.
Saunders said that despite the cost savings from running on Revio, the test is still about twice as much as short-read WGS. Sequencing reagents for one Revio run alone cost around $3,680; on a per-genome basis, Pastinen said sequencing costs are approximately $1,400, compared to approximately $600 for short-read WGS. For comparison, Rady charges about $8,000 to $9,000 for its rapid WGS test, including interpretation.
Studies of rapid WGS have shown that much of the cost savings come from reduced hospital stays. Whether more information from long reads can offset the higher baseline costs and longer time to results remains to be seen.
Bamshad noted that his own study won't have data on HiFi genomes for a while. "We have nearly 50 families sequenced with Oxford Nanopore [sequencing]; however, we won't have 50 with Revio until next year."
But the most complex scenarios, "where you need to rule out with high confidence repeat and imprinting disorders," provide an opportunity for real costs savings when using long-read WGS, Saunders said.
Bamshad further suggested that the test's value "will probably differ, based on the cohort that's being looked at. … Diagnostic rates are, in part, widely variable. Even two cohorts of the same condition can be different."
Pastinen's hypothesis is that long-read WGS will provide more impact as a first-line test. "Once a patient leaves without a diagnosis, it's unlikely there will be another test and families can go through years of testing," he said. "I do think there's potential that we see more of an impact than our research lets us suggest."
Past research has required a specific type of parental consent, he explained. In the clinical utility study, Children's Mercy will be able to offer a different level of disclosure. "Since the samples are not used in research dissemination, individual cases are not studied," he said.
The LDT could also offer value as a second-line test. Children's Mercy works with several medical centers around the country to receive cases that have gone unsolved, including the University of Nebraska Medical Center, University of Utah Health, Indiana University Health, and New York University Langone Health.
"I think they're excited to see if we're going to recover diagnoses after failed rapid WGS," Pastinen said. As such, he has applied for funding to analyze "a few hundred" samples with long-read sequencing as a second-line test. An ongoing pilot study is looking at 50 babies where rapid WGS did not lead to a diagnosis.
Overall, Pastinen said he would be surprised if diagnostic yield improved more than 10 percent, but suggested his utility study would also look at how overall care could be impacted through a single test.
"That might actually be much more surprising to the community and to us, in terms of what does a single comprehensive test do early in genomic testing versus a traditional approach," he said.