Oct 20, 2016
By Filipa Lynce, MD
Lombardi Comprehensive Cancer Center, MedStar Georgetown University Hospital
Claudine Isaacs, MD, FRCPC
Lombardi Comprehensive Cancer Center, Georgetown University
The traditional model by which an individual was identified as harboring a hereditary susceptibility to cancer was to test for a mutation in a single gene or a finite number of genes associated with a particular syndrome (e.g., BRCA1 and BRCA2 for hereditary breast and ovarian cancer or mismatch repair genes for Lynch syndrome). The decision regarding which gene or genes to test for was based on a review of the patient’s personal medical history and their family history. With advances in next-generation DNA sequencing technology, offering simultaneous testing for multiple genes associated with a hereditary susceptibility to cancer is now possible. These panels typically include high-penetrance genes, but they also often include moderate- and low-penetrance genes. A number of the genes included in these panels have not been fully characterized either in terms of their cancer risks or their management options. Another way some patients are unexpectedly identified as carrying a germline mutation in a cancer susceptibility gene is at the time they undergo molecular profiling of their tumor, which typically has been carried out to guide treatment choices for their cancer.
Throughout the past several decades, we have witnessed tremendous advances in our knowledge of evaluating and treating patients with germline mutations in hereditary cancer syndromes, with studies clearly demonstrating the feasibility and clinical utility of genetic testing. Perhaps most importantly, studies have provided convincing evidence that implementing prevention strategies in some instances prolongs the survival of mutation carriers. For example, for unaffected women who carry a BRCA1 or BRCA2 mutation, risk-reducing salpingo-oophorectomy results in a significant reduction in all-cause mortality (3% vs. 10%; hazard ratio [HR] 0.40; 95% CI, 0.26–0.6), breast cancer-specific mortality (2% vs. 6%; HR 0.44; 95% CI, 0.26–0.76) and ovarian cancer–specific mortality (0.4% vs. 3%; HR 0.21; 95% CI, 0.06–0.8) when compared with carriers who chose not to undergo this procedure.1 Additionally, Markov modeling suggests that a 30-year-old healthy BRCA1 mutation carrier would gain 0.2 to 1.8 years in life expectancy with risk-reducing salpingo-oophorectomy and 0.6 to 2.1 years from risk-reducing mastectomies.2,3 Given these findings, genetic testing for hereditary cancer syndromes has now become part of standard practice.
As set forth in the Evaluation of Genomic Applications in Practice and Prevention (EGAPP) initiative and further supported by the recent ASCO policy statement on testing for genetic and genomic cancer susceptibility, a number of criteria must be considered when evaluating existing or emerging genetic tests.4,5 These criteria include analytical and clinical validity, clinical utility, and the associated ethical, legal, and social issues. In the context of genetic testing, analytic validity refers to the accuracy and reproducibility by which the assay detects the presence or absence of a mutation. Clinical validity focuses on whether the test accurately and reproducibly predicts the clinically defined disorder. Clinical utility can be defined as the evidence that a genetic test results in improved health outcomes typically based on early detection or prevention strategies, and the test’s usefulness and added value to patient management decision making. For genetic testing, particularly for moderate-penetrance genes, clinical utility remains the fundamental issue.5 The EGAPP framework is key when evaluating the utility of genetic testing for hereditary cancer syndromes. Failure to meet some of these criteria forms the basis for many concerns regarding the current clinical actionability of multigene panel testing.
Single/Limited Gene Testing
For well over a century, it has been recognized that some families harbor a hereditary predisposition to a variety of malignancies. In 1913, Warthin described a kindred known as Family G, in which he noted an aggregation of endometrial carcinoma along with gastric and colorectal cancer.6 This family, among others, formed the basis of the initial descriptions of hereditary nonpolyposis colorectal cancer syndrome, now more commonly known as Lynch syndrome. Similarly, astute clinicians recognized other hereditary cancer syndromes such as Li-Fraumeni and Cowden syndromes and hereditary breast and ovarian cancer based on the cancer phenotype of the family.7-12 By the mid-1990s, linkage analyses and other studies resulted in the ability to pinpoint individual genes associated with some of these hereditary cancer syndromes.13,14 These included the identification of BRCA1 and BRCA2 associated with hereditary breast and ovarian cancer; MLH1, MSH2, MSH6, PMS2, and more recently EPCAM associated with Lynch syndrome; FAP with familial adenomatous polyposis; and TP53 with Li-Fraumeni syndrome.
However, it became apparent that many families with striking histories consistent with either a hereditary colorectal or breast/ovarian cancer syndrome are not found to carry a mutation in one of the mismatch repair genes associated with Lynch syndrome or in BRCA1/2. For example, studies indicate that a mutation in a mismatch repair gene is found in approximately 40% to 80% of families that meet the Amsterdam I criteria and only about 5% to 50% of families meeting the Amsterdam II criteria.15 Similarly, only about 5% to 10% of unselected patients with breast cancer and 20% to 25% of patients with hereditary breast cancer are found to carry a deleterious BRCA1 or BRCA2 mutation. Additionally, a number of other genes associated either with rare high-penetrance syndromes or a more moderate penetrance were identified.16,17
Based on these findings, the general paradigm of testing evolved whereby the more common genes such as BRCA1 and BRCA2 were tested first, and, if negative, sequential testing for additional gene(s) was performed if the patient met criteria for testing for other syndromes. This process had both advantages and disadvantages. In terms of advantages, the genes tested in this setting typically have well-described cancer risks and often have established management guidelines. Additionally, through the pretest counseling process, patients undergoing this testing have had the opportunity to fully consider the benefits, risks, and limitations of testing in their particular situation. In terms of disadvantages, such testing is less comprehensive than multigene testing, and, if performed, sequential testing is quite time-consuming and costly.
There has been a dramatic shift in the genetic testing landscape over the past several years, in large part because of two major factors. The first is the development of next-generation sequencing, a high-throughput approach to DNA sequencing that allows for massively parallel sequencing of multiple genes more efficiently and at a lower cost than the traditional Sanger sequencing methods. The second is the Supreme Court decision in 2013 for Association for Molecular Pathology v. Myriad Genetics, which invalidated many patents restricting BRCA1/2 testing. Very shortly after the ruling, many companies and some academic institutions announced they would offer BRCA testing in addition to the existing genes on their multigene panels.18,19 As a result of these two factors, offering relatively rapid turnaround times for multigene testing in a reasonably affordable manner became feasible.
A number of studies have evaluated the utility and impact of multigene testing in a variety of settings. The key questions that must be addressed revolve around the clinical utility or actionability of the findings from such testing, namely (1) the numbers of patients who are found to have a deleterious mutation in a gene for which cancer risks are known and management strategies exist, (2) patients who are found to have a mutation with uncertain cancer risks and/or no evidence-based recommendations for management, and (3) the rate of detection of variants of uncertain significance (VUS). The rate of VUS varies between 3.3% and 42%, and many patients were reported to have two or more VUS. The VUS rate is still high in some reports, but it is expected to fall in the near future because of the rapid accumulation of data from multigene panel testing.
Single/Limited Gene Versus Multigene Panel Testing
Several factors guide the decision to pursue testing of a single gene or a finite set of genes associated with a particular syndrome versus multigene panel testing. These include (1) the characteristics of the proband’s personal and family history, (2) an individual’s preferences and tolerance regarding the possibility of ambiguous results, (3) insurance-related issues, and (4) the rapidity with which results are needed. A publication in The New England Journal of Medicine in 2015, authored by experts from the United States, United Kingdom, the Netherlands, Germany, Australia, and Canada, thoughtfully reviewed the issues that must be addressed when considering multigene panels.20 Additionally, ASCO released a policy statement on genetic and genomic testing for cancer susceptibility to reflect the impact of advances in this field.5
Single/limited gene testing remains an excellent option when the clinical features, such as the patient’s personal and family history, are strongly indicative of a particular syndrome associated with a single or finite set of genes. This approach allows for a focused and comprehensive pretest evaluation in which individuals have an opportunity to more fully consider the impact of testing for a particular gene or set of genes. Additionally, such testing minimizes the likelihood of detecting a VUS or a deleterious mutation in a gene with limited clinical information.
Multigene panel testing is an appropriate option when the family phenotype is not suggestive of a single specific mutation and one or more hereditary cancer syndromes are in the differential. Additionally, panel testing is often considered if more focused initial testing is negative (e.g., BRCA1/2 testing followed by multigene breast/ovarian panel). Multigene panel testing has a number of advantages and potential disadvantages (Table 1). The advantages include gains in efficiency both in terms of cost and time. Such testing also would result in a more comprehensive assessment of the genes that could account for the cancer phenotype in the family. Finally, pragmatically, in this era of multigene panel testing, it is unclear if an individual could obtain insurance coverage for repeat testing if the initial, more limited testing results were negative.
In terms of disadvantages, as described previously, multigene panel testing has a higher rate of detection of VUS. Individuals undergoing testing must be fully informed of this possibility before testing and counseled on the interpretation of such a result. Furthermore, it is important that an individual undergoing panel testing understands it is possible that a high-penetrance mutation in an uncommon or rare gene may be identified, even in the absence of a classic presentation of the associated syndrome. Consequently, aggressive interventions may be recommended, such as consideration of prophylactic gastrectomy if a CDH1 mutation was found, even in the absence of gastric cancer in the family. At this moment, it is unclear if the cancer risks for patients identified through panel testing without features of the associated syndrome are the same as quoted in the literature because of ascertainment bias. Moreover, laboratories have varying methods by which they assure the analytic and clinical validity and the clinical utility of the variants they report. Expertise in this area is required to ensure accurate interpretation of the clinical significance of the findings reported. Given the panoply of testing options, this expertise is also critical to guide the choice of which test to order and from which laboratory. These issues further underscore the importance of ensuring that patients undergo pre- and post-test genetic counseling by well-trained professionals, as endorsed by ASCO and the National Comprehensive Cancer Network.21
Next-generation sequencing has introduced substantial complexity and promise in the field of cancer risk assessment. Although multigene panel testing provides a more comprehensive and efficient approach to testing an individual for a hereditary susceptibility to cancer, the information obtained can be challenging to interpret. Furthermore, many of the genes included in multigene panels have not been fully characterized either in terms of their cancer risks or management strategies. In many cases, single/limited gene testing remains a very appropriate testing option. Presently, we live in an era in which our technical capabilities have outstripped our medical knowledge. A strong and continuous partnership among clinicians, individuals with genetics expertise, and laboratory geneticists is critical to bridge this gap.
As to the detection of incidental findings on tumor sequencing, more research is clearly necessary to better clarify how to approach this complex area. Until such time, as stated by ASCO, it is critical that individuals undergoing tumor sequencing be fully apprised of the possibility, benefits, risks, and limitations that such testing could uncover unanticipated mutations in cancer susceptibility genes.
- Domchek SM, Friebel TM, Singer CF, et al. JAMA. 2010;304:967-75.
- Schrag D, Kuntz KM, Garber JE, et al. JAMA. 2000;283:3070-72.
- Grann VR, Jacobson JS, Thomason D, et al. J Clin Oncol. 2002;20:2520-29.
- Teutsch SM, Bradley LA, Palomaki GE, et al. Genet Med. 2009;11:3-14.
- Robson ME, Bradbury AR, Arun B, et al. J Clin Oncol. 2015;33:3660-7.
- Warthin AS. Arch Intern Med (Chic). 1913;12:546-55.
- Li FP, Fraumeni JF Jr. Ann Intern Med. 1969;71:747-52.
- Li FP, Fraumeni JF Jr. J Natl Cancer Inst. 1969;43:1365-73.
- Lloyd KM II, Dennis M. Ann Intern Med. 1963;58:136-42.
- Le Dran H. Mem Acad Chir (Paris). 1757;3:1-54.
- Smithers DW. Br J Cancer. 1948;2:163-7.
- Broca PP. Traité des tumeurs. Paris: P. Asselin; 1866.
- Hall JM, Lee MK, Newman B, et al. Science. 1990;250:1684-9.
- Nakamura Y, Lathrop M, Leppert M, et al. Am J Hum Genet. 1988;43:638-44.
- Lynch HT, de la Chapelle A. N Engl J Med. 2003;348:919-32.
- Antoniou A, Pharoah PDP, Narod S, et al. Am J Hum Genet. 2003;72:1117-30.
- Nathanson KL, Wooster R, Weber BL. Nat Med. 2001;7:552-6.
- Azvolinsky. J Natl Cancer Inst. 2013;105:1671-2.
- Cook-Deegan R, Niehaus A. Curr Genet Med Rep. 2014;2:223-41.
- Easton DF, Pharoah PDP, Antoniou AC, et al. N Engl J Med. 2015;372:2243-57.
- Hall MJ, Forman AD, Pilarski R, et al. J Natl Compr Canc Netw. 2014;12:1339-46.