Stomaching Multigene Panel Testing: What to Do About CDH1?

In their timely Commentary in this Journal issue, Bryson Katona and colleagues (1) address the emerging problem of unexpected findings following multigene panel testing (MGPT) for “cancer” genes. They provide a daunting example of a recurrent scenario in the cancer genetics clinic: MGPT of genes for which justifiable guidance is lacking. The story portrays a 43-year-old female proband with no personal cancer history but a family history concerning for Lynch syndrome. Genetic testing uncovered a likely pathogenic variant in CDH1, a gene in which damaging variation increases the risk for diffuse gastric cancer and lobular breast cancer (2), yet there was no history of either in the proband or family. Cascade testing (3) uncovered additional family members harboring the likely pathogenic CDH1 variant, and the proband and other carriers elected to undergo total gastrectomy. Particularly worrisome, the MGPT pointed to an unexpected cancer-predisposition gene, CDH1, for which there is considerable uncertainty as to proper management. Katona et al. (1) wisely emphasize the need for careful pretest genetic counseling for individuals undergoing CDH1 testing. However, the presence of CDH1 on many MGPT raises a challenging question: Should patients be offered the opportunity to opt out of CDH1 testing?

Today, ordering MGPT usually includes a set of genes for which the patient does not meet criteria and for which the clinical risks are uncertain. In this regard, much of the content of MGPT can have limited clinical utility, questionable cost-effectiveness, and substantive unforeseen consequences. Opting out of genetic testing (in this case, requesting that CDH1 not be included) is technically consistent with established medical principles of “do no harm,” respect for autonomy, and the right not to know but is nowadays practically challenging. The authors’ recommendation fits with current practice on the return of genetic results. The American College of Medical Genetics and Genomics (ACMG) maintains a list of 59 genes that, if variation is found to be pathogenic or likely pathogenic, should be returned to patients (unless they opt out), regardless of original testing indication (4). Pathogenic variation in the genes on the ACMG list are linked to well-established, serious health risk for which specific interventions and management strategies are available that can attenuate risk. Notably, CDH1 is not on the ACMG-59 list. Because few clinical options are available to manage gastric cancer risk in an individual with a pathogenic CDH1 variant (unproven endoscopic screening or gastrectomy) and the cancer risks in an individual in the absence of a concerning personal or family history are unknown, Katona et al. (1) argue that a reasoned discussion before testing is warranted. Alternatively, this could be addressed during the consent process by clearly stating that the involved MGPT tests a number of genes for which the full clinical picture is as of yet unknown.

Most commercial and academic MGPT include genes for which we cannot provide clear guidance and that blur the linebetween clinical and research testing. This raises the question of whether we should ask for consent to test genes for research simultaneously. For many of these genes, it will take time to accrue sufficient data and careful annotation to suitably interpret the information in a way that health-care providers can discuss with patients and their families. Thus, MGPTs effectively conflate diagnostic and research testing in the same offering. Testing beyond well-understood high-penetrance cancer-predisposition genes for which the clinical utility is clear requires more detailed pretest counseling and consent, as well as considerably more time and resources.

One could reasonably extend the recommendation to exclude CDH1 in MGPT to modalities such as clinical exome and genome sequencing. The factors that should prompt a second thought on testing CDH1 (uncertain penetrance, limited understanding of biology, lack of understanding of the sequelae of precursor lesions, and morbidity of intervention) can of course be applied to other genes. There is precedent for this in the practice of return of Alzheimer-related pathogenic variation in APOE in exome sequencing (5).

The counterargument to this selective exclusion of a particular gene raises the specter of “genetic exceptionalism.” Genetic test results are certainly not unique in their ability to provoke anxiety and infuse uncertainty into clinical care. Although we generally agree with the arguments in Katona et al. (1), exclusion of CDH1 (or other genes) may not be a long-term solution for a host of reasons. These include the increased time and labor burdens imposed on genetic counselors, also measured in additional cost, not necessarily covered. Laura Hercher wisely stated that the right to know and the right not to know are two sides of the same autonomy coin (6). In the context of return of genetic test results, our field has tended to focus on the latter. Increasingly, studies suggest that individuals cope well with genetic test results, even when unanticipated (7–9).

In the age of the “Wild West” of MGPT, selective reporting has become the focus of considerable debate. Although the informed exclusion of CDH1 (or other genes) from MGPT (or an exome or genome report) may be justified, the expansion of genetic testing services and technologies is swamping the capacity to selectively limit what is available to patients and consumers. CDH1 testing exemplifies the challenge of “genetic contextualism,” in which the current state of knowledge and specific circumstances merit consideration during disclosure of genetic information (10). This is why it might be propitious to consider dual consent for clinical indication and for research up front.

Further-outcomes research is needed to clarify cancer risks in the setting of a lack of personal and family history of cancer. CDH1 is the tip of the iceberg for genes found on commercial clinical tests in which precise risk estimates are insufficiently available, and evidence exists that patients are being mismanaged based on uncertain genetic test results (11). Particularly in the setting of CDH1, more research is needed to determine progression of foci of signet ring cell carcinoma to fulminant diffuse gastric cancer to inform recommendations around gastrectomy in those with low-to-moderate penetrance phenotype. The ClinGen consortium (12) is tailoring ACMG variant interpretation guidelines (13) for specific genes, including CDH1 (14), but this will take time and new data, some of which could be aggregated from consented MGPT. Additional guidance is under development by ClinGen on the interpretation of lower penetrance variation. Multigene genetic testing in the form of next-generation sequencing panel—and increasingly genome and exome testing—continues to proliferate, in no small part driven by commercial pressures through the development and marketing of ever more comprehensive and complex testing offerings. We must prepare our patients for the future of genetic testing while keeping at the forefront the guiding principles of clinical utility, the balance of benefits and harms, and informed consent. It is this latter point that necessitates careful planning for collection of clinical testing when guidance is clear, as well as when more data are required to eventually interpret and provide guidance based on sufficient evidence.

Funding

This work was supported by the Intramural Research Program of the Division of Cancer Epidemiology and Genetics of the National Cancer Institute, Bethesda, MD.

Notes

The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US government. The funder had no role in the writing of this editorial or the decision to submit it for publication.

Dr Stewart provides telegenetics services for Genome Medical, Inc, in accordance with relevant National Cancer Institute policies. There are no other disclosures for other authors.