NEW YORK – Almost four years after the ruling in favor of Quest Diagnostics in one of the most notable genetic testing lawsuits in recent years, there is still variability in how labs and physicians view their responsibility to recontact patients when a genetic variant's classification changes, despite growing standardization in how these classification changes are made and recorded.
Quest was sued in 2016 by Amy Williams, whose son's sample was tested by Athena Diagnostics, which was later acquired by Quest.
The lab initially reported back a variant of unknown significance in a gene associated with an epilepsy syndrome but doctors reportedly never saw the reclassification report sent by the lab. It was thus not considered in the patient's care, and he later died.
Williams argued in her complaint that Athena was negligent and that the reclassified result might have changed the course of her son's disease. However, the US District Court for the District of South Carolina eventually found the lab not responsible.
Eight years later, there has been some consensus building on best practices and responsibilities for recontacting physicians and patients regarding variant reclassifications, but the nature of the complex web among laboratory practices, medical records technologies, rapidly evolving medical and genetics literature, and physician practices has ensured that outcomes continue to vary.
Clinical genetics expert Heidi Rehm, codirector of the program in medical and population genetics at the Broad Institute and chief genomics officer in the Department of Medicine at Massachusetts General Hospital, said that one significant challenge is that patient recontact remains a notably "resource-intensive process."
"I think the challenge right now is that … the kind of infrastructure [labs] have," Rehm said.
An example, she said, is the GeneInsight software that she and colleagues developed for their own medical genetics program in the Mass General Brigham medical system in Massachusetts. GeneInsight is designed to alert physicians or genetic counselors when the knowledgebase is updated with new information about patients' variants.
"We spent many years building that, with massive investment, and it works great. But it only works for the physicians who order tests from my lab." Rehm said.
"I can say what the ideal is, but that's based on [a lab] having pretty expensive systems to manage this stuff, so there's a fine line between saying that labs should reinterpret [and] issue updates if variants change, and then actually getting the infrastructure necessary," Rehm said. "That is, by definition, going to be variable because it's dependent on resources."
Ambry Genetics Chief Medical Officer and Laboratory Director Elizabeth Chao agreed, though she stressed that variation in practices among leading labs reflects these kind of implementation challenges and not necessarily differing opinions on the value of appropriate patient recontact.
"If you think back to when genetic testing started to grow in the mid-2000s, there were no recommendations, there were no best practices," Chao said. A trainee at the time, she remembers first encountering an explicit recontact policy at Myriad Genetics, which dominated hereditary cancer testing at the time under later-dissolved patents.
"I think that the overwhelming trend over time has been that both labs and clinicians have really learned to appreciate the value in this, which has led to … where I would expect quality labs offering genetic testing services … to offer some form of recontact when a test result changes," Chao said.
"There's some variability, obviously, in policies, but I would argue it relates more to staffing and resourcing, rather than necessarily a difference in opinion about how important this is," she added.
For example, before joining Ambry, Chao ran several smaller academic labs. "We didn't have the right resources we needed to build the types of programs that we run at Ambry, but it wasn't because we didn't believe they should happen," Chao said. "We used what we had, and we reached out for the most critical cases."
Current practices
Chao described Ambry's current system for reclassifying variants and recontacting ordering physicians as an interplay between standards for how and when reclassifications are prompted and how that translates into new test reports and physician outreach.
As has become the standard in the field, the company classifies the variants it identifies when it sequences a DNA sample into one of five categories: benign, likely benign, uncertain, pathogenic, or likely pathogenic.
But while the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) have issued guidance documents on how to classify variants, there isn't the same consensus for when and how often laboratories should review the classification of a particular variant.
"For our practices, we have time-based guidelines for how long variants are allowed to stay in those categories before they expire," Chao said. "For example, variants that we categorize as benign, we actually have no expiration on those. We don't purposely go back and ever look at them again."
For other categories, the timelines can be as frequent as every six months.
This variant-specific timeline system is coupled with proactive evidence review, whereby if a new piece of evidence comes out — whether from a population frequency database like the Broad Institute's GnomAD (also headed by Rehm) or a published functional study — the company takes that evidence and queries its own database to see if there are any affected variants.
"We have an internal standard of practice for how we're proactively addressing the literature," Chao said.
She added that Ambry is also heavily involved with collaborative variant resolution groups, through which it gets notified of variants that are somehow in conflict between member labs, which can also drive reclassifications.
At Ambry, in the case of a reclassification, there is an automated process that identifies patients affected and generates a list of cases to recontact.
"Although we do have EMR integrations, this piece of it is not automated for us internally, and we have different processes," Chao said. For example, if the shift is from likely pathogenic to pathogenic, the company does reach out to ordering physicians, but at a lower priority than if something shifts to or from a VUS.
"It's not just a notice that goes out," Chao said. "It's a follow-up conversation that's needed to really discuss why that happened and make sure they're aware of it, and that is on us."
The same is true for genetic counseling. Ambry doesn't mandate it but tries to support physicians in providing it to patients where it can.
Sometimes recontacting an ordering physician is difficult or even impossible, she admitted, in which case, "we do the best that we can."
Myriad Genetics, also a continued leader in the space, especially in hereditary cancer genetics testing, described its own protocols for patient recontact as a "lifetime commitment," which, much like Ambry's policy, involves monitoring scientific and clinical data for all its tested genes and variants.
"When a variant in any gene on our hereditary cancer panels is reclassified as clinically actionable, we will issue an amended report to the healthcare provider," Susan Manley, Myriad's senior VP of medical services, said in an email.
The company does not send amended reports for reclassifications that do not impact the clinical care of patients, for example, a shift from uncertain to likely benign.
Aspects of the company's literature review are, again, automated, but Manley said that Myriad also does a manual review before anything is reclassified or amended results are sent.
In the recontact process, a member of the company's customer service team reaches out to the ordering provider to verify their receipt of the amended report. Its medical services team handles any specific follow-up questions about the reclassification.
"Follow-up with the ordering healthcare provider may not be successful in cases where a provider has retired, moved to a different healthcare practice, or if the healthcare practice is no longer in business," Manley added, saying that in these cases, Myriad does attempt to notify patients directly that there is new information regarding their test results and ask them to complete a provider designation form. They can also opt to have a copy of their report sent directly to them.
Like Ambry, Myriad doesn't provide genetic counseling as part of its test orders, but it does have an in-house team that can answer questions from outside counselors, other providers, or patients.
Neither of the two largest lab companies in the US — Laboratory Corporation of America, which recently moved to acquire hereditary cancer and other genetic testing assets from longstanding field leader Invitae, and Quest, which fielded the lawsuit that has brought attention to these procedures over that last decade — replied to a request for comment on their current protocols for patient recontact.
Shared responsibility
Among the guidance that does exist regarding patient recontact is a set of statements developed by ACMG that provides "points to consider" for laboratories and clinicians regarding patient recontact after revision of genomic test results.
Rehm said that ACMG's overall take on this has been that all parties involved have a responsibility. However, because labs themselves are in the best position to communicate when variant information changes, they are the most responsible for doing that.
But clearly, it's not all up to the lab. Physicians who receive updated reports are the ones who have to track down the patients, which is not always straightforward. "That's another headache," Rehm said.
Regarding efforts to standardize variant reclassification protocols, "that's easier because that's science … and we have made tremendous progress there," said Chao, who chaired one of the working groups behind the ACMG's points to consider. "But there are ongoing challenges with the practicality of implementing these [other] policies. Patients move, patients change doctors, no one wants to be left holding the bag when a result changes."
Having been involved in writing one of the statement documents, Chao said she views the content as largely general and nondirective, noting that "I don't think that they solve the problem [of standardization], but they did at least give us a source to point to."
Chao also highlighted work by Wendy Chung, a medical geneticist currently at Boston Children's Hospital, that has included the convening of a number of focus groups bringing together not only genetics professionals but also lawyers, ethicists, insurers, genetic counselors, and other physicians.
Drawing on interviews with these various stakeholders, Chung and colleagues have published several papers outlining the need for development and dissemination of best practices for variant reclassification and patient communication.
Meanwhile, in a commentary published in the European Journal of Human Genetics last March, two members of the departments of pediatrics and bioethics at the Hospital for Sick Children at the University of Toronto — Michael Mackley and Lauren Chad — argued that the equity implications of a shared model of responsibility in patient recontact have been "insufficiently discussed" and that to ensure equity, recontacting practices "must be universal" — something Rehm and Chao both made clear is decidedly not the case.
The risk, Mackley and Chad wrote, is that although responsibility for recontact is intended to be shared, it ends up being "overwhelmingly patient-driven" in practice, especially in Canada's publicly funded healthcare system.
This presents an equity issue, as patients who most frequently take this proactive role tend to be those with social advantage, such as having higher levels of education and health literacy, speaking the dominant language, and belonging to the majority ethnic group.
Contextual differences
Systems for patient recontact after variant reclassifications may also be easier to harmonize in the context of certain diseases versus others. In rare disease and pediatric sequencing cases, for example, the relationship between physicians and patients or families can be especially close and enduring. The knowledgebase among clinicians regarding variant reclassification is deeper and the potential impacts on patients more dramatic than in other areas, including oncology.
According to Chao, while the feedback that Ambry gets from patients and families following a reclassification notice varies a bit across different disease areas, it is overwhelmingly positive. "Most of our patients and families are looking for an answer: a reason for either their medical condition or why so many people in their family have cancer," she said.
"They're seeking to understand that, so although reclassification is maybe not the news that they've hoped for, we have found that more data, more information, seems to be incredibly valuable to patients and families, many of whom are still on some sort of diagnostic odyssey or trying to understand their family history better," Chao added.
With clinicians, the variability in experience seems greater, she said.
"In the setting of rare disease or pediatrics, where we're working a lot of times with medical geneticists or pediatric neurologists who have deep, long-term care relationships with these patients and families, they understand the process, they've been through it before, and they're very positive when they're provided with new information," said Chao.
In contrast, it can be much more challenging for physicians in a primary care practice or even in an oncology practice, as these same long-term relationships might not exist. "If they're not following patients year after year, or maybe if they're not doing high-volume orders of genetic testing, this may have never happened in their practice before, [whereas] I think you'd be hard pressed to find a medical geneticist who hasn't delivered reclassification results," Chao said.
A survey of oncologists last year on the subject of patient recontact in hereditary cancer gene testing showed a variety of views and practices in the field. Publishing in JCO Precision Oncology, investigators reported on responses they received from 634 individuals including 349 oncologists and 285 genetic counselors. Interestingly, both groups largely responded that all reclassified variants, even those that don't affect clinical management, should be returned to patients, in contrast to numerous clinical lab policies.
Regarding the frequency of recontacting patients with reclassified results, almost half of the surveyed genetic counselors reported recontacting patients often, compared to only about 13 percent of oncologists.
Rehm said that there are a couple of factors that play into the difficulties faced in variant reinterpretation, which then trickle down to the struggles faced in developing policies for recontact.
One, she said, is the prior probability of a positive result for a given test. "When you're doing cancer panel testing, most of those tests have a very low prior probability of positive finding," Rehm said. "And so, just statistically speaking, any variant that's not likely pathogenic or above is highly unlikely to ever become pathogenic."
The vast majority of variant changes are moving to likely benign or benign, "and that's just a huge overhead with little utility to the patient," Rehm added. "Assuming you are appropriately not using the VUS information for that patient's care, when that changes to benign, it removes a headache, but you shouldn't have done anything with that to begin with."
Sachdev Thomas, a medical oncologist at Kaiser Permanente's Vallejo Medical Center, said in an email that in his experience, downgrading VUS to benign or likely benign occurs more often — 3 percent to 5 percent of the time — than the opposite, which occurs about 1 percent of the time.
He said that he has not personally seen a reclassification occur that would affect patient management or risk for family members, though the center does have a genetics department that tracks this more broadly.
Rehm and colleagues have been working on other ways of improving the baseline of variant classification rules to potentially help overcome these obstacles, including a future publication that she said is expected to support updated guidelines for variant reporting, based on data from her and other labs' experience subclassifying VUS.
Her lab, along with three others, examined the implications of subdividing VUS into different tiers depending on their evidence level. Rehm said that the lowest class of VUS — dubbed "VUS-low" using new terminology that she and her colleagues developed — are virtually never reclassified to either pathogenic or likely pathogenic, suggesting they could be completely left off of results reports or at least not featured prominently and instead be included in supplementary materials.
Changes like this could reduce some of the stickiness that oncologists and other practitioners face when they field unactionable variant reclassifications, allowing for more uniformity as standards hopefully evolve for labs and doctors in their shared responsibility to inform patients.