NEW YORK (GenomeWeb) – At a meeting held in Washington, DC, last week, genetics and legal experts urged an independent committee to study the growing liability risks for labs when interpreting genomic variants and recommend best practices for sharing data and working collaboratively to resolve classification discrepancies.
The National Academies of Sciences, Engineering, and Medicine's Committee on Science, Technology, and Law (CSTL) — chaired by David Baltimore, a Nobel laureate and Caltech president emeritus, and David Tatel, a judge for the US Court of Appeals for the District of Columbia Circuit — heard from genetics and legal experts on the liability issues raised in Williams v. Quest/Athena, in which Amy Williams is suing Quest Diagnostics and its subsidiary Athena Diagnostics of misclassifying a genetic variant for her infant son, Christian Millare, who died eight years ago.
At the meeting, lab director Heidi Rehm and science policy expert Robert Cook-Deegan discussed the case as a stark example of why there needs to be a better system for test providers and researchers to share information on the relationship between variants and diseases. Moreover, the classification of genetic variants — whether they are likely benign, benign, of uncertain significance, likely pathogenic, or pathogenic — may change with accumulating evidence, but standards are lacking on how labs should reinterpret results and update physicians and patients, they said.
As genomic sequencing is getting cheaper and more accessible to patients, labs are identifying a lot of variants that are rare in the population, often seen in a single patient. These are challenging to classify, and genetic test providers, under pressure to perform testing quickly and cheaply, are sometimes interpreting variants incorrectly.
There also is a recognition, across government, academia, and industry, that there needs to be more transparency and collaboration when classifying the innumerable variants that might occur in 20,000 genes, but there isn't much consensus on how disparate players in the life sciences space should come together to tackle this — and not everyone is eager to do so.
"There's a really hard slog that we're going to have to go through," said Cook-Deegan, a professor at Arizona State University's School for the Future of Innovation.
Rehm, director of the Laboratory for Molecular Medicine at Partners HealthCare Personalized Medicine, told the committee that even with standards for interpreting variants, genetics professionals may come to different conclusions about the pathogenicity of a variant.
In Williams v Quest/Athena, Williams is suing because she believes that based on the published literature and according to the lab's own classification criteria her son's variant should have been classified as pathogenic, instead of a variant of uncertain significance, which is how Athena interpreted it back in 2007. Quest and Athena have asked the court to dismiss the lawsuit on statute of limitation grounds.
Rehm evaluated the variant at issue in this case using the most recent interpretation guidelines from the American College of Medical Genetics & Genomics and the Association for Molecular Pathology, and pointed out how using the recommendations a genetics expert might decide that there isn't enough evidence to make a definitive classification. But she noted that using the same guidelines other professionals might decide otherwise (see related story for in-depth coverage of these issues).
"We can all appreciate that the practice of medicine doesn't represent complete concordance across professionals," she said. "But we can dramatically reduce that discordance through data sharing and consensus building."
Rehm has been urging genetic test providers and researchers to submit to ClinVar, an archive of genotype-phenotype relationships launched by the NIH three years ago. ClinVar now contains records on more than 158,000 unique variants with interpretations from close to 600 submitters. Genetic sequencing firm Illumina announced last week that it contributed 95,000 genetic variants to ClinVar, making it the top submitter.
However, those not participating in ClinVar seem to dislike that fact that ClinVar reveals the classification discordance between labs. Some perceive this as increasing their liability.
In ClinVar, 17 percent of variants interpreted by more than one lab are classified differently. But Rehm told the committee that when nine labs got together to evaluate a set of variants according to their own criteria and by ACMG/AMP's guidelines, they were able to talk through the discrepancies for 22 percent of the variants and reduce the discordance to 5 percent.
Although there currently is no requirement that labs publicly share data, Rehm has been working with payors, regulators, and lab-accrediting bodies with some success to encourage test providers to submit to ClinVar.
So far, she has been able to impress upon the US Food and Drug Administration and a few payors the importance of sharing variant data. Medicare contractor Palmetto GBA has recommended labs submit to ClinVar in a recent coverage policy, and Aetna is requiring labs submit BRCA1/2 variants to ClinVar in order to ink new in-network contracts.
Meanwhile, the FDA wants to recognize public genetic variant databases that meet certain standards, so developers of next-generation sequencing tests can reference the information in the repository to establish the validity of variants gauged by their assays, instead of doing clinical studies. "FDA could probably use some external help on thinking about what those rules would look like," Cook-Deegan told CSTL members. According to Rehm, ClinVar will likely apply for FDA recognition.
However, any FDA effort to regulate genetic tests invariably gets bogged down in debates about whether the agency has the authority to regulate tests developed and performed in laboratories, which are traditionally overseen by Centers for Medicare & Medicaid Services under the Clinical Laboratory Improvement Amendments. And so, there are also efforts to tweak the existing system of oversight so labs will be more inclined to submit to public databases.
For example, Rehm and others involved in ClinGen — which in collaboration with ClinVar aims to build a genomic knowledge base that can be used in patient care — have been in discussions with lab accrediting body the College of American Pathologists about the possibility of requiring that labs share variant information in a public database as a condition for accreditation.
But based on a statement from CAP, which accredits thousands of labs in the US and abroad, it seems not much progress has been made. A spokesperson told GenomeWeb that the organization is not involved with ClinVar "as it is outside the scope of CAP." Although CAP does have representation in a ClinGen working group, the spokesperson said that "nothing from the working group's discussion has been brought back formally to CAP for further consideration at this time."
The field could benefit from CSTL's advice on how to address these challenges, Rehm and Cook-Deegan agreed, particularly at a time when the FDA, NIH, and national projects like the Cancer Moonshot and the Precision Medicine Initiative are encouraging data sharing. "I think regulatory bodies need to enforce this," said Rehm, noting that CAP, CMS, and FDA could all play a role. "CSTL could help with that enforcement in this space," she added.
They noted that since the US Supreme Court three years ago struck down the ability to patent genetic sequences as they exist in the body, new business models have emerged with genetic testing firms competing on services but not trying to monopolize test markets for specific genes. The test providers that are voluntarily submitting to ClinVar, among them GeneDx, Invitae, and Ambry Genetics, say they are doing it to be more transparent, provide more accurate results to patients, and improve the field's understanding of variants.
"We need to create an environment where labs compete on services and they don't compete on patient safety," Rehm said. "That data that is the basis for accurate interpretation of [variants] … needs to be shared as a standard in patient safety."
CSTL, which includes members with expertise in science and law, has previously held workshops, collected data, and issued reports on topics, such as intellectual property held by universities, Avian flu research, and management of complex scientific evidence in legal cases. After the meeting last week, Anne-Marie Mazza, senior director of the committee, told GenomeWeb that the CSTL hasn't yet decided how it wants to address the topic of liability for diagnostic labs and is considering the full range of issues.
This is one of two articles on a recent National Academies of Sciences, Engineering, and Medicine's meeting that explored liabilities for diagnostic labs. See related story.