NEW YORK – The American College of Medical Genetics and Genomics last month published a set of points that clinicians and genomics labs should consider when presumed germline variants are flagged during tumor sequencing.
With the increasing use of large next-generation sequencing panels in cancer care, labs are detecting variants that patients may have potentially inherited and that may be informative for their future cancer risk, as well as their treatment. However, since the goal of tumor testing is to identify molecular features of patients' tumors that may be targeted by treatments, presumed germline findings in this context may be unexpected for patients and their doctors, and oncologists may not be prepared to address the clinical significance of such findings. Moreover, within the lab industry, there isn't a standardized way of handling germline findings from tumor testing, and every lab is dealing with this differently.
Recognizing this is an issue of increasing importance in oncology, an ACMG work group put forth a "points to consider" document in Genetics in Medicine to encourage best practices among doctors and labs and spark further discussion. Based on a review of the literature and their own experiences, the experts laid out for labs the potential for germline findings with different testing approaches, described features that can increase the likelihood that patients harbor germline variants, and discussed challenges related to germline variant confirmation, interpretation, and reporting.
For clinical professionals, the work group discussed test selection, the importance of consenting patients for germline findings, and genetic counseling. The paper also includes several case studies to highlight, for example, the reimbursement challenges with confirmatory germline testing in this context, and the importance of this information to patients and their families.
Marilyn Li, director of cancer genomic diagnostics at the Children's Hospital of Philadelphia, and Douglas Stewart, a senior investigator within the National Cancer Institutes' Division of Cancer Epidemiology and Genetics, Clinical Genetics Branch, are members of ACMG's professional practice and guidelines committee, and co-leading authors of the document. In a recent interview, they discussed the growing importance of this issue to ACMG and to the broader cancer care community. Below is an edited transcript of the interview.
Why did the ACMG decide to publish this "points to consider" document on this specific topic now?
Stewart: There are several reasons. First, tumor sequencing is increasing in volume and increasingly important in the practice of oncology. Second, there is increasing availability of therapies based on germline variation, such as PARP inhibitors for breast cancer. Third, around 10 percent of pediatric cancer and varying percentages of adult cancers are attributable to germline variation, and identifying that variation is key to managing risk in patients and families through surveillance protocols and cascade testing. And last, a significant number of patients are missed based on existing screening recommendations. We need to take advantage of every opportunity and tool we have to identify people at risk.
You outline three testing scenarios in the paper that are happening right now: tumor-only testing; tumor/normal testing where germline findings are subtracted out; and tumor/normal testing where labs are reporting only a subset of germline findings associated with cancer risk. What is the most common method right now?
Li: We reviewed a large number of publications and survey results about current practices with regard to next-generation sequencing of tumors. We summarized that information into three major approaches and each has its advantages and disadvantages. For example, if you do tumor-only analysis, the advantages are that it's lower cost. The turnaround time could be shorter because you don't have to wait for a second specimen to get to the lab. You don't have to go through a lot of patient consent, because you're not dealing with germline variants. But certainly, there are disadvantages, for example, you might miss a clinically relevant germline variant.
For tumor/normal paired analysis with subtraction, it offers greater accuracy in detecting somatic variants [and it] streamlines the data analysis process as germline variants are all subtracted, so that you only need to curate somatic variants. It will miss clinically relevant germline variants if any, and has a higher cost. For the hybrid approach where tumor/normal paired analysis is done with a specific group of cancer predisposition genes analyzed, you'll know whether a variant in one of these genes is germline or somatic right away when you get the data, but you have to sequence two genomes, compared to one.
The most common approach is tumor-only analysis. There are many reasons for that, but it comes back to the cost.
The cost of next-generation sequencing comes from two components: the wet lab where you prepare the library and do the sequencing, and the dry analysis of the data that you get from sequencing. The larger the panel [of genes being tested], the higher the cost. Outside of the actual sequencing, the amount of time you spend on variant curation can also increase costs. Most labs offer small-to-medium size NGS panels [with] 50 to a few hundred genes, [and] the number of [detected] variants is manageable, especially for those hotspot mutation panels.
Then, the cost to run one sample, as opposed to two, could be double, especially for those labs who are running specimens at or near capacity … There are also other issues, such as getting access to normal samples, and consenting the patients.
ACMG does not mandate which approach laboratories should take. We want to raise awareness that presumed pathogenic germline variants can be inferred even if you do tumor-only analysis. For labs that don't have the capacity to do germline confirmatory testing or have the resources to do pre- or post-test genetic counseling, they should refer those patients to labs or centers that do have this capacity. Each lab should decide based on their capabilities. ACMG would recommend that each laboratory clearly describe its testing spectrum and limitations to avoid any false expectations or misunderstandings from referring physicians.
Are labs being transparent about their testing capabilities and what they are reporting and not reporting when it comes to potential germline findings from tumor testing?
Li: In this "points to consider" document, ACMG did not evaluate if labs are being transparent about detecting or reporting presumed pathogenic germline variants. Our job is to promote medical best practices. This document recommends specifically that all laboratories should clearly describe their test spectrum, methodologies, and limitations regarding detecting and reporting germline variants from tumor testing.
You noted that cost is one of the main reasons labs aren't reporting out potential germline findings from tumor testing. But in your own practice, what are you seeing from a cost/benefit standpoint? Even with the added costs, are pathogenic germline variants being missed at a significant enough rate to warrant that labs make these investments from an ethical and downstream healthcare cost standpoint?
Li: Studies have shown that a significant fraction of germline pathogenic variants are missed using established guidelines. We've seen in recent publications that pathogenic germline variants are being identified using tumor testing. One of our goals through the publication of this document is to raise awareness that these clinically significant variants can be missed if we're not mindful about them … That could lead to more resources being poured into this effort. Having said that, the laboratories should still decide based on their own resources how to handle the germline variants identified through tumor sequencing.
Stewart: I agree with that. It's been sobering [to see] in the publications noted in our paper the number of patients that are missed who may have a pathogenic germline variant even if appropriate guidelines are followed for genetic testing. Germline reporting from tumor/normal sequencing is another tool in the toolbox to identify as many people as possible at risk … Another important point in the last paragraph of the paper is that the development of practice guidelines related to the detection and return of these presumed germline pathogenic variants should really include all stakeholders, including geneticists, oncologists, pathologists, payors, governmental and private [entities], and patient advocacy groups. We need to work together to figure out the best ways to handle this opportunity.
Some of the most interesting aspects of this paper are the case studies included to highlight how germline findings can impact patients. There's an example in the paper where tumor-only testing in a child with rhabdomyosarcoma identified a TP53 missense variant and there was reason in this case to think that this patient should receive confirmatory testing to see if this variant occurred in the germline. But the variant allele frequency from tumor-only testing fell just shy of the insurers' lower limit for coverage for confirmatory analysis. To what extent are payor denials the rate-limiting step to reporting germline findings after tumor testing, and how common is it for insurers to use variant allele frequency as a threshold for coverage?
Li: To my knowledge most insurers do cover cancer predisposition testing or confirmatory testing if it is deemed clinically relevant. Physicians may need to send letters to justify that, the insurance company may have [additional] questions, and then [doctors] may have to write follow-up letters answering those questions. In most cases, if you can answer all the questions and [show] that testing is justified, then the testing is covered.
Stewart: That's been my experience, too. It can require lots of follow-up work in the form of phone calls and letters by the physicians, and many times, from the genetic counselors. I'll also point out that ACMG is a tireless, longstanding advocate for the reimbursement of genetics services. This is a battle that is ongoing.
Li: It's our hope that this document will also help with reimbursement.
There are different sets of points to consider laid out in the paper, and one set is for the physicians. Importantly, one of the points that clinicians should consider is "the likelihood of an underlying cancer predisposition syndrome." That suggests that may be not all oncologists are considering the possibility of germline findings when they do tumor testing. Is that the case, and if so, why is it?
Stewart: I don't want to speak for oncologists since I’m not one. But in the paper we address this and encourage clinicians of all types to develop professional relationships with a geneticist or a genetic counselor to help with these considerations both in the pre-test and post-test spheres. On top of that there is increasing recognition of the importance of germline findings in managing risk in patients and families, and for treatment.
An aspect of this question hits upon the issue of rare diseases. There are thousands of rare diseases, but in aggregate they're not rare. A given practitioner may or may not see a particular rare disease in their career. We hope through this "points to consider" there is a recognition that geneticists do know rare diseases and can serve as a resource with our genetic counseling partners to help clinicians manage these patients.
The paper makes several suggestions for ensuring that germline findings don't fall through the cracks. Labs are recommended to not just report the findings to a patient portal but also have more interaction with physicians to communicate the significance of findings. And healthcare providers are recommended to develop a referral network of genetics experts who can more appropriately guide patients. Are there any positive indications suggesting that more labs are engaging with doctors around communication of germline findings and that these referral networks are forming and being utilized?
Li: There are indications from recent publication and webinars suggesting that more laboratories are engaging with doctors of different disciplines around communication of germline findings.
ACMG recommends that genomic results should be delivered to patients by qualified healthcare professionals to maximize their understanding of test results and avoid any potential confusion. We, at Children's Hospital of Philadelphia, have recently published our institutional pipeline for communicating test results of presumed germline pathogenic variants. We serve as a referral center for quite a few hospitals and institutions, and this was a joint effort with genomic laboratories, cancer predisposition programs, and cancer center oncologists.
This document also discussed different mechanisms for returning genomic test results to referring physicians and patients. Using multi-disciplinary tumor boards have been very popular [as a mechanism for returning results and discussing those results], not only within major academic centers, but also [among] community oncologists. I have participated in some of those tumor boards and found them very helpful.
In the paper, the ACMG emphasizes getting patient consent to report back germline findings and letting patients know that they might still find out about clinically significant germline variants incidentally from tumor only testing. In current practice, are healthcare providers appropriately consenting patients for the possibility of germline findings in the context of tumor testing?
Li: For tumor-only testing, patient consent is not necessary, but there may be state-specific requirements and laws ... In general, if a germline variant is suspected, the patients and their parents [in the case of pediatric patients] should be referred to a geneticist or to a genetic counselor for further evaluation and that would determine if a confirmatory test in necessary. If it is, then the patient will go through genetic counseling and will be consented before testing.
For those labs offering hybrid tumor/normal analysis, the patient should be informed in those cases that there is the possibility of germline findings, even though the main goal is identification of somatic mutations. If likely pathogenic or pathogenic germline variants are identified, and confirmed, then … those patients should have post-test genetic counseling.
You mentioned earlier that the role of the ACMG is to promote best practices, and noted that in this paper, the group is not saying that germline findings must be reported as part of tumor testing. But if in the future you do form a more collaborative group involving different stakeholders, do you think ACMG might take a more definitive stance on whether labs should be reporting germline findings when they do tumor genetic analysis?
Stewart: This is a complicated topic and there are lots of stakeholders involved. This particular document is called a "points to consider" but it's also a type of ACMG document. It is a way to get a conversation started. It is a way for ACMG to provoke conversation that may eventually lead to additional documents that support a position one way or the other. In the spirit of that, in the last paragraph of this paper, we call for more study, more evidence, and more discussion among everybody involved to capitalize on this huge opportunity of identifying germline variation in tumor sequencing so that it benefits as many people as possible.