NEW YORK – Stanford Medicine's Center for Inherited Cardiovascular Disease transitioned this week to clinical whole-genome sequencing for all of its in-house gene panel tests covering different classes of inherited conditions.
By implementing a whole-genome backbone for these assays now, the center believes it can continue to offer comprehensive, clinically validated testing, while also making it easier to expand its offerings over time and to reflex to broader analyses if an initial test fails to identify the cause of a patient's disease.
Moreover, the move should help accelerate the discovery of new pathogenic variants and potential polygenic risk factors by amassing more whole-genome sequence data linked to clinical cases.
The launch is the first of several that Stanford plans to make this year to expand its clinical genomics footprint, with future steps including the opening of a new preventive genomics program and the launch of polygenic risk score and pharmacogenomics testing for both cardiovascular disease and primary care patients.
Euan Ashley, director of the Center for Inherited Cardiovascular Disease and its larger clinical genomics program, said that although the institution has been using whole-genome and exome sequencing in its undiagnosed diseases program for some time, its cardiovascular genomics clinic had previously sent out all gene panel testing.
"I think starting from when genomes first became relevant to clinical medicine, that's probably when we started thinking about … [looking] towards a future where we would use a singular genome backbone to do testing across the healthcare enterprise," Ashley said.
"We've been working from that day to the point that we could really apply this more broadly."
Across institutions, most genomic testing for inherited cardiovascular conditions is panel-based, with some panels only covering a small number of genes of interest and others a broader set. "To our knowledge, very few places have rolled out a genome backbone," Ashley said. "Some use the genome as default in place of exome, but I think in terms of it being the [foundation] for all the testing, this is the first step in that direction for us here and I think very exciting because of all the extra things that the genome can bring beyond just the capture of specific genes."
The falling cost of sequencing has certainly been a tipping point, Ashley added, along with growing recognition of the advantages of collecting more comprehensive data. "For one, there is the ability to immediately reflex, so if you don't find the answer in the original gene, you can reflex from a panel of 50 genes to the exome equivalent" without the need to perform additional sequencing, he said.
Greater coverage also aids variant calling, and allows the interrogation of intronic regions that standard capture would miss and where, increasingly, clinically relevant variants are known to occur.
Stuart Scott, laboratory director for the Stanford clinical genomics program, added that starting from a whole genome also allows for efficient updating of panels as new evidence emerges.
"As new genes are discovered in cardiovascular diseases, we can very quickly go back and open up that window and revalidate those regions and incorporate them into the clinical test," he said.
Looking to the future, a whole-genome backbone could eventually support panel-based tests in other areas of medicine at Stanford, he added. "We now have this infrastructure in place, so forthcoming panels will be so much easier to set up now."
Scott said that the whole-genome assay has been validated for single nucleotide variants, insertions/deletions, and structural variants. For cardiovascular applications, Ashley and colleagues have created six subpanels to read out, covering approximately 200 genes in total and ranging from just four genes for the smallest panel to more than 100 for the largest.
At Stanford's CICD "we take referrals from a very broad range of the west coast of the US and beyond that, internationally as well, for patients with Mendelian inherited cardiovascular disease … and we see about 30 new families every week, so this will be offered to all of those patients who come in," he said.
Kiran Musunuru, director of the genetic and epigenetic origins of disease program at the University of Pennsylvania's Perelman School of Medicine and lead author of an American Heart Association statement on genetic testing for inherited cardiovascular diseases that was published last year, said in an email that he wouldn't expect the implementation of whole-genome sequencing to significantly increase the number of patients who receive genetic diagnoses in the short term, but that it's clear that it can help accelerate scientific discovery of new genes and mutations that will improve the care of patients with inherited cardiovascular diseases in the longer term.
Ashley agreed, saying that the rate of discovery of new genes has dropped in recent years, even though as many as 50 percent of patients, depending on the condition, don't get an answer from existing panels.
"One of the real benefits of doing this much more broadly is that … once we start to accumulate hundreds or thousands of patients, we can start to look across those genomes and potentially make new gene discoveries that we wouldn't be able to make from a single individual," he said.
"Although this is a clinical tool, I think it's the discovery aspect that will also move this along and can really help the clinical delivery, eventually, by adding the genes with good evidence."
Stanford's move reflects growing recognition of the value of broader genomic profiling, Musunuru added, with other complementary initiatives emerging, such as Geisinger Health's move to offer large-scale sequencing to patients, or a recently launched clinical study of whole-genome sequencing by institutions in New York.
"The downside of routinely doing whole-genome sequencing," he added, is that it can turn up more variants of uncertain significance: mutations with unknown consequences for patients' health.
"It could make genetic counseling and management of patients with concern for inherited cardiovascular diseases that much more challenging," he cautioned.
According to Ashley and Scott, Stanford's approach of sequencing the whole genome but limiting its clinical use to established gene panels directly reflects these caveats.
Variant classification and reporting for the new genome-backed cardio panels follow ACMG guidelines and the core content is centered around clearly "clearly causal genes" with an option to release data on genes with more moderate evidence if a clinician requests it, Ashley said.
In about three months or so, Ashley, Scott, and colleagues are planning to add polygenic risk score (PRS) testing for these same patients, with scores that can be calculated from the same whole-genome backbone.
This will parallel the launch of a separate, new preventive genomics pilot program where the same cardiovascular PRS will be offered to patients in a primary care setting. Because those individuals are not having their genomes sequenced at Stanford, the university is partnering with commercial sequencing firm Personalis to perform low-pass genome sequencing on which the PRS can be run.
"The algorithm will be the same, but it will be applied and called either off the high-depth genome or a low-pass [version] for the broader preventive primary care setting," Ashley said.
Unlike the cardio gene panels, the Stanford team did not develop this PRS component in house. "We had the option to kind of make our own bread, but in this case, because we were interested in this as a clinical rollout, not a research-based program, we elected to work with Genomics PLC," Ashley said.
"From the lab perspective, our focus is on validating the actual analytical piece, which is actually all of those millions of sites that drive the algorithm, and then validating the algorithm itself so that our scores are equivalent to what we expect," Scott said. "We have to do that both on our clinical whole genome … and then navigate doing the same from the low-pass sequencing, as well."
Along with PRS, Stanford Medicine is also preparing to launch pharmacogenomic testing for both the preventive genomics effort and for reporting from the CICD's whole-genome backbone.
Despite the fact that pharmacogenomic variants, especially for cardiovascular drugs, have been known for many years, adoption of PGx into clinical practice has been slow. But Scott said that recent progress in standardization and guidelines bodes well for more systemic implementation in coming years.
According to Ashley, the ability to implement PGx preemptively is especially key. "If you're a physician going to write a prescription, even if you've got all the information at your fingertips that this is a potentially relevant test, if you have to then send a test and wait to get a result back … that just doesn't work with the way healthcare is organized. But if that information is already in the chart … if it's already there in front of you, then I don't know a physician who wouldn't take advantage of that additional information."
Recent controversy over laboratories offering direct to consumer PGx tests with questionable content put a "bit of a stain" on the field in recent years, Scott added. "But now … is really the ideal time for implementation because we have so many more resources. It's not just the laboratory saying 'we've found a variant in the drug-response gene.' There are actually clinical societies that are publishing peer-reviewed and very rigorous clinical practice guidelines for a subset of gene-drug pairs … and even the FDA has [now] produced their pharmacogenomics tables," he said.
To support its planned PGx addition, the Stanford team has been working with Pacific Biosciences to incorporate long-read sequencing into its platform alongside the short-read whole genome approach, with the goal of launching that test later this year when the preventive genomics pilot program begins.
"The nice thing about having the whole genome is that a lot of pharmacogenomic variants are not in coding regions, so we definitely get exposure to those sites, which is fantastic," Scott said. "But one of the challenges is variant phasing … so for that, we're working on the long-read sequencing mechanism to actually interrogate the pharmacogenomics panel."
The plan is to launch a defined panel, but to evolve it as evidence accumulates or changes for less well-defined gene-drug interactions emerge, he added.
Musunuru said that other healthcare systems, including UPenn, are mirroring Stanford in exploring both PGx and polygenic risk testing in preventive care, though he cautioned that there are still questions that need to be answered about the clinical utility of these emerging technologies.
"There's a lot of energetic debate right now about the most appropriate use … in clinical practice," Musunuru said, "with one big area of concern being that polygenic risk scores developed in one group of people (for example, people of European ancestry) don't necessarily work well in other groups, raising issues of equity."
In their AHA statement, he and his coauthors highlighted that recent data on PRS for complex cardiovascular diseases have suggested that "patients with extreme scores, that is, in the top few percent of the population, have an increased risk of disease, severalfold higher than that of the population average," reaching an equivalent risk elevation to that conferred by some monogenic disease genes.
However, "whether such information is actionable and can meaningfully inform patient management remains to be determined," they wrote.