The known unknown: the challenges of genetic variants of uncertain significance in clinical practice

As genetic testing technology has advanced, allowing scientists to obtain much of the raw data from our DNA, their ability to interpret these data has struggled to keep up. The result is the ubiquity of variants of uncertain significance (VUSs): findings from genetic testing for which the clinical significance is currently unresolved. What to do when these results are found is a problem that has vexed laboratories, clinicians, and patients alike. In this paper, I focus on the issues raised by VUSs in clinical practice, and suggest paths forward. The completion of the Human Genome Project in 2003 was widely heralded as the dawn of a new age of medicine—which, in many ways, it has been, enabling stunning advances in genetic testing and research. However, the ‘book of life’ is written in a language whose alphabet we can read, but whose vocabulary and grammar we do not yet fully understand. As a result, we have the technical ability to report how the ‘words’ of genes are spelled, but still often lack the interpretive ability to say what they mean. Genetic testing is performed clinically for a variety of reasons. Diagnostic testing is done to determine whether a patient with symptoms has a suspected condition, or to determine what condition they have, if their symptoms are non-specific. For example, a child with seizures and dysmorphic features might receive genetic testing to determine whether these features are associated with a known genetic condition. This testing can sometimes provide an explanation, predict recurrence risk, or inform clinical management. Predictive testing or genetic risk assessment is usually performed for people with a family history of a condition for which they have not yet presented symptoms. One example is testing of the BRCA1/2 genes, mutations in which can dramatically increase the risk of breast and ovarian cancer, in people who do not yet have cancer. These types of tests can sometimes lead to clinical interventions such as surgery or enhanced screening, or be useful in life planning. Reproductive testing includes carrier screening, which can inform parents of the risk of having a child affected with a recessive genetic disorder, and prenatal testing, which can determine whether a pregnancy is affected by a genetic condition. These decisions may be used to inform reproductive decision making, including whether and how to become pregnant or whether to continue with a pregnancy, or to help families prepare for the birth of a child with a genetic condition. No matter the context of testing, results generally fall into three categories. ‘Positive’ results usually provide a diagnosis or risk information. ‘Negative’ results, where no relevant genetic variation is found, can sometimes rule out a diagnosis or reduce a risk assessment, or in other cases simply be uninformative. In between positive and negative falls the gray area of the VUS. Every human's DNA sequence—with the usual exception of monozygotic twins—is unique. While some of the difference between individuals is responsible for disease, the majority of it is benign. When DNA testing is performed, the data obtained can be compared to a ‘reference sequence’, and differences from that sequence identified. However, it is not always possible to predict with confidence whether a difference, or variant, found is actually deleterious or just attributable to genetic diversity. Even rare variants in genes known to be associated with disease are now understood to often be benign, contrary to the early expectations of researchers. Variant classification, the process of determining whether a DNA variant causes disease, is in clinical practice generally the province of genetic testing laboratories and relies on multiple lines of evidence. These include whether a variant has been previously seen in individuals with a condition and has not been seen in those without it; what its effect is predicted to be on the protein produced by the gene; and others. While some variants can be confidently predicted using these criteria to either be pathogenic (disease-causing) or benign, in many cases, pieces of evidence are missing or conflict with each other. In these situations, the variant is often classified and reported as a VUS. Recent advances in next-generation sequencing technology that have allowed the rapid and inexpensive sequencing of numerous genes at one time (panel testing), of the regions of the genome that code for proteins (the exome), or the entire genome, have increased the yield of genetic testing and made it more accessible—but have also greatly increased the number of VUSs encountered in clinical practice. Simply put, the more places you look, the more likely you are to find a VUS. The wonderful promise of expanded testing has, in practice, run into the frustrating reality of a greater burden of uncertain results. Another increasing source of VUSs is chromosomal microarray (CMA), a test that looks at larger sections of DNA and can discover copy number variants (CNVs), or pieces of the genome that are missing or duplicated. Many CNVs are well-described causes of genetic conditions, or clearly disrupt genes in a way that can be predicted to cause disease. Still others have been seen in healthy individuals and are almost certainly benign. Many, however, also fall in that vexing middle ground that qualifies them as VUSs. In 2000, the American College of Medical Genetics (ACMG) first issued guidance for genetic testing laboratories on the classification of sequence variants, and established six categories of variant classification, ranging from pathogenic to benign, with VUSs in the middle.1 This guidance was revised in 2007 and again in 2015.2 Similar guidance has been issued for the classification and reporting of CNVs.3 Even with the detailed criteria laid out by ACMG, the process of variant classification is still more art than science, and laboratories often differ in how they classify variants—though efforts are underway to share data between labs and make classification more consistent.4 ACMG guidance for clinical laboratories recommends that labs report VUSs in genes related to a clinical indication.5 One relatively unique characteristic of VUSs, in comparison to other types of ambiguous medical test results, is that while the result itself may remain static, its meaning is often resolved over time, as more data are gathered. Laboratories routinely reclassify VUSs—they may be ‘upgraded’ to pathogenic or ‘downgraded’ to benign (the latter being more common). When this occurs, laboratories generally issue amended reports to healthcare providers to, in turn, disclose to their patients. While there has been a great deal of attention given to how labs classify and report VUSs, this article focuses on the challenges presented once a VUS result reaches a clinician's, and then a patient's, hands: in terms of clinical management, counseling, and impact on patients. VUSs are now a daily fact of life in genetics practice, but little guidance is available for how best to approach them, and limited data are available on how they are affecting medical practice and patient well-being. The receipt of VUS results by clinicians raises a number of questions. Should they be disclosed to patients, and how should the patients be counseled? Should they inform clinical management? What follow-up studies should be done? What happens when a variant is reclassified? These questions are troublesome enough for genetics professionals, but are even more challenging for clinicians without specialist training in genetics, who are increasingly encountering genetic test results—and therefore, VUSs—in their practice. The ACMG’s variant classification guidelines address the issue of management as follows: ‘A variant of uncertain significance should not be used in clinical decision making. Efforts to resolve the classification of the variant as pathogenic or benign should be undertaken. While this effort to reclassify the variant is underway, additional monitoring of the patient for the disorder in question may be prudent’.6 While somewhat helpful, this statement still leaves great deal up to the clinician's judgment. In contrast to research, where a great deal of debate exists about what results should be disclosed to participants,7 the general consensus in clinical practice is that once VUSs are disclosed to providers, they should in turn be disclosed to patients. However, not all clinicians agree. One study on BRCA1/2 found that some genetic counselors would consider not disclosing VUS results under certain circumstances.8 Clinicians also express a great deal of disagreement on how much detail they would go into in disclosing a VUS.9 Beyond what clinicians say they would do when surveyed, it is difficult to know how often, and in what manner, VUSs are actually disclosed in practice. In some cases, as noted by ACMG, the significance of a VUS can be resolved through further testing, but it is unclear how much of that work a clinician should be expected to pursue. One common type of follow-up is testing of family members to determine whether variants are shared by other affected or unaffected individuals. This can be time-consuming and labor-intensive, and relies on family cooperation and consent. Further complicating matters, labs sometimes pay for these follow-up tests, but not always, and insurance usually does not. Another possible follow-up is pursuing assays that may be able to determine the functional consequences of a gene variant—however, these assays are usually the province of research scientists and not generally clinically available. Following up on VUS results can often blur the line between patient care and research, and with a shortage of genetics professionals,10 the time required for these efforts can be prohibitive. In terms of clinical management, ACMG notes that VUSs on their own should not change how patients are treated—although the caveat about the prudency of ‘additional monitoring’ leaves some wiggle room regarding conservative changes, such as screening. However, increasingly, medical professionals without genetics expertise are ordering or encountering genetic testing in their practice, and there is evidence that they may be failing to understand the import of VUSs or acting inappropriately in response to them. A striking example can be found in studies of surgery decision making after receiving VUS results for breast cancer susceptibility genes. Bilateral mastectomy (BLM) is a dramatic intervention that has been shown to have no survival benefit for the majority of women; the possible exception is women with genuine pathogenic variants in BRCA1/2. In a recent study of surgery decisions made by women with early-stage breast cancer, researchers found a surprisingly high rate of BLM in women with otherwise average breast cancer risk after learning of a VUS in a breast cancer susceptibility gene, and a significant number of breast surgeons indicated they managed patients with VUS the same as those with pathogenic variants11—strikingly different from how most genetics professionals would manage these patients, or from the ACMG’s guidance. A case of a major intervention being recommended by a cardiologist in response to a VUS in a cardiac disease gene has also been reported.12 It is unknown how prevalent this type of mismanagement is, but it is likely to only become more common. One more issue raised by VUSs in clinical practice is that of variant reclassification and recontact. If variants are reclassified by a laboratory, the laboratory is generally understood to be responsible for issuing a revised report to the clinician. The process of then getting this information to patients is not always straightforward. Variants are often reclassified many years after the original test,13 and providers may have changed practices, or patients may have out-of-date contact information on file. How much of the legwork of tracking down old patients clinicians can be held responsible for, and whether a ‘duty to recontact’ exists in general, is another question that remains up in the air.14 The ACMG’s most recent policy statement on the duty to recontact is from 1999 – well before VUSs became so ubiquitous—and states that clinicians are responsible for re-contact for patients for whom they provide ongoing care (typically not the majority in genetics practice); otherwise, the responsibility lies with the patient's primary care physician, or the patient themselves, to check in periodically.15 It seems impractical to expect primary care physicians to regularly contact a specialist for years to check for variant reclassification, but equally impractical to expect overworked genetics professionals to track down patients from years ago. This question is less troublesome in the case of a downgraded VUS result, but the importance of an upgraded result making its way to a patient could be quite significant. There are no data available on how often reclassification information makes it to providers and then patients, but it seems likely that a great number of patients never find out that their variants have been reclassified. A case currently making its way through the courts, while focused on the conduct of a genetic testing laboratory, also vividly illustrates the challenges of clinical response to a VUS. Williams v Quest/Athena is a lawsuit filed against Quest Diagnostics and its subsidiary Athena Diagnostics by the mother of a young boy who tragically died of a seizure disorder at the age of 2.16 In 2007, Christian Millare's genetic testing revealed that he had a variant in a gene called SCN1A, some mutations in which are known to be associated with the seizure disorder Dravet syndrome. Athena reported it as a VUS, despite the existence of two papers presenting the case of another epileptic patient with the same variant. This failure to diagnose Christian with Dravet syndrome was highly consequential because certain anti-epileptic drugs are known to worsen seizures for patients with Dravet, and unfortunately, Christian continued to be prescribed these drugs while his seizures worsened and he eventually died. Amy Williams, Christian's mother, obtained a copy of his test result in 2014 with the help of a genetic counselor and learned the variant had been reclassified; in 2015, Quest/Athena issued an amended report classifying the variant as pathogenic but citing the same evidence used in the original report. Williams has sued Quest and Athena, alleging negligence in how they dealt with Christian's results. While the case has been filed against the laboratory, it also raises questions about the responsibility of clinicians with regard to VUSs, and where the responsibility of the lab ends and the clinician's begins. To begin with, Williams claims that the SCN1A test results were never even disclosed to her, which is now widely regarded as a clinician's duty. It is also unclear what additional actions Christian's doctors should have taken in response to the VUS finding. For one the same papers available to Athena could have been found by Christian's doctors with a of the after receiving his could have follow-up testing on Christian's parents to help the of the variant or they could have even a VUS related to Dravet syndrome as a for certain falls the of ACMG’s for further but not its There as no guidelines as to which of these actions Christian's doctors should have but Quest and Athena have to the lawsuit by alleging negligence on the of the This case the import of how VUS results are and the for guidance for It is to be the case to be filed these clinicians disclose a VUS result, the question then of how this information is being understood and by patients. a VUS result has been as patients in VUSs are difficult for patients to and can decision the focused by receiving one of these results may also cause be to patients, and about VUS in a way that and the negative of this Genetic information is and challenging for patients to VUSs are difficult in this The of an uncertain is at with of the generally genetics as than and a VUS result lies in the It can also be difficult for patients to that a variant could be when it is found in a gene that a or family history of disease. efforts to uncertain results can often fall and patients and or lead to In and general genetics, genetic testing is often done to to find a diagnosis for a patient with the to find an for these patients is a at a VUS this can be frustrating for patients and this the or This for may and of one study of parents who uncertain results from found that many of the VUSs as for their Similar in have been found in the cancer One study found that while of genetic with a VUS in BRCA1/2 being that it was an result, of them it as a genetic to This that even with counseling, there may still be a between what patients are being about a VUS and what they are Another study found that of VUS was more common in patients with the importance of this information in a way that is to The from a VUS can be greatly for patients. on the impact of VUS results are but there is some evidence at in the case of receiving an uncertain result may lead to and than receiving an even cancer genetics professionals care to them Another context in which receiving a VUS result can be challenging is the prenatal is increasingly used but not for with a VUS result unique challenges in this where limited data and time make it more difficult to the of a variant, and where decision making, such as pregnancy is highly and uncertain results from testing has been found to be for they report from not having enough time to the information they they to make a decision and a high burden of with one it In an age where prenatal genetic testing and are increasingly these are likely to more and more patients, with no to making them less this evidence of and most studies have found that the majority of patients to receive VUS results, even in prenatal with many citing an of is as with so many in what people and how they when they it can be different There are limited data on and how often, patients are about the of receiving VUS results before receiving genetic A study found that but not counselors patients about the of VUSs in for However, genetic testing is increasingly by providers, who are less likely to VUSs before and the shortage of genetics professionals leaves less time for even in genetics This raises questions about and a lack of may patients in a in terms of and with uncertain results. variant classification in also of in genetic The classification of variants relies on an of the genetic variation in a and the data that from people of As a result, patients of other have been found to be more likely to receive VUS results from and cancer This that all of the challenges presented are likely to patients, a frustrating and of The to the existence of VUSs is studies of genes and variants, and data with variant classification and reduce should be focused on the of data from currently in to address and The of these data is a however, and with a great deal of genetic variation being rare and even unique to we are likely with VUSs for the In the of genetic that VUSs is dramatically patients and A possible policy change that could help resolve some VUSs is for insurance to family testing, which is currently insurance have had a difficult time to the that testing other people can actually be relevant to a patient's medical In the laboratories could in some up the as the of a variant is also for them. The approach to VUS can sometimes the of a from the lab to the clinician on to the who is the one who has to with the of the who is generally to understand One way out of this by genetic counselor is for laboratories to the to not report VUSs, but to keep them on and clinicians if the variant is reclassified to While this has its the case of Christian one if by clinicians, of a VUS in SCN1A could have Christian's It is that a VUS out to be and even that it would have such a dramatic impact on clinical management, but this be against the to of VUS given the challenges of re-contact disclosing VUS results clinicians and patients to do their own research and track down variant reclassification if that information to be The fact that most patients to receive VUS results should be taken into as them these results could be or for patients to whether they would to about VUSs could also be These may be however, an from the approach to VUS reporting seems VUSs to a of best clinical for with them is genetics professionals, and most for the professionals who are increasingly encountering VUSs in their practice. The for genetics professionals has been the of a great deal of it is that this includes information on VUSs to the consequences of and guidelines and that into the increasing ubiquity of VUS and VUS reclassification are The practice of the ordering for a genetic test when a variant is reclassified may to be medical professionals could some of VUS reclassification contact who responsible for this information efforts may also be to patient contact information over time, even after a patient is from genetics There are likely to be that can the negative impact of VUSs on patients and When patients receive regarding the of VUS findings and their lack of clinical they may be less likely to be or when they receive research is on different and and how they impact and in response to VUSs. The reality of VUSs is not much as clinicians and patients might it Many challenges to be on a with and between clinicians, and patients. further research and guidelines help make the through and eventually out of genetic less and