Focusing on the Future of Genetic Testing in Oncology

John M. Carethers, MD, MACP, discusses recent developments in multi-gene panel testing.

Germline genetic testing is essential in order to identify optimal treatments for patients with cancer, as well as detecting inherited mutations via cascade testing that could affect family members, according to John M. Carethers, MD, MACP, who emphasized that improvements to genetic testing technology and testing costs has increased not only the accuracy of, but access to these assays.

The technology in sequencing has moved from the old gels to capillary to ChIP [chromatin immunoprecipitation]-based, and has revolutionized the way we approached it. The depth of [genetic testing] coverage [has evolved],” said Carethers. “Sequencing technologies totally revolutionized this [process].

He added, “There are some unusual situations in which additional technologies have to be used to figure out some of the ones that typical ChIP technologies don't fully explain. That has markedly changed the way we approach [testing] these days.”

In an interview with OncLive® during the 2020 Institutional Perspectives in Cancer (IPC) webinar on Precision Medicine, Carethers, a professor of Internal Medicine and Human Genetics at the University of Michigan, discussed recent developments in multi-gene panel testing.

OncLive®: How are predictive and somatic genetic tests being utilized in clinical practice?

Carethers: In terms of germline testing, the benefit is knowing which disease you carry, and that information can also spread to other family members to understand whether they [are at an increased risk of getting a cancer diagnosis]. Sometimes, at least in my experience, [germline testing] does alleviate some anxiety. Some people get more anxious once they know they have a germline mutation, but in general, it does at least explain the reason why they're seeing certain diseases in the family. That’s the general benefit for germline testing.

The benefit of somatic testing is knowing the type of mutations that occur in the tumor; there may be a therapeutic drug or compound that is in current use that could benefit the patient. For instance, I had a patient with unresectable esophageal cancer. She was dying and her esophagus was almost completely obstructed with the tumor. She had a feeding tube put into her stomach and lost a lot of weight; she was literally counting out the days until she died. With some thought, we decided to take a sample of the tumor and do somatic testing.

She had some mutations that weren’t typically found in esophageal cancer, and we did have drugs [to treat her]. She was actually put on those drugs and the tumor shrunk dramatically to the point that she could eat again, she gained weight, and she lived another 5 years. Normally, she wouldn't have lasted more than a few months. The benefits of somatic testing is understanding the genetic makeup of the tumor in which you might be able to use some compounds that exist to benefit the patient. That’s the real goal of somatic testing.

There is an unusual situation for somatic testing, as well. For instance, in colon cancer, we know about Lynch syndrome, but there is also a Lynch-like syndrome. In Lynch-like syndrome, there is no germline [mutation], but the tumor has 2 somatic mutations of a mismatch repair deficient tumor. They can look like a Lynch syndrome tumor, and maybe even behave a little bit like a Lynch syndrome tumor, but they're really not caused by a germline mutation. Sometimes, somatic genetics can help us understand tumor genesis as well as ways to treat the tumor.

What changes have we seen recently in multigene panel testing? How are test results interpreted and how do they help guide treatment strategies?

There are patients who will walk in with the classic phenotype and then there are patients walking in who don't have the classic phenotype, yet they carry that mutation in the same gene. Multigene testing allows us to account for phenotypic variation.

Someone may walk in with colon cancer, the next person in the family might walk in with endometrial cancer, and the next person in the family may walk in with a skin tumor, but they all line up with the same mutation in Lynch syndrome. However, if you saw the skin tumor first, would you have thought of Lynch [syndrome]? [What about] if you saw the endometrium or the colon cancer? It depends on the specialty and the type of disease presentation they show up with. In many cases, though, the disease could be subtle.

For instance, there was a family I followed, which comprised the grandmother, mother, and daughter. The grandmother, who was well into her late 60s, had a Lynch syndrome mutation and got her colon removed appropriately. The mother was in her 40s with no cancers, but the daughter who was 21, developed colon cancer. It looked like it skipped a generation, yet, they all carry the same mutation. There's phenotypic variation, even with this exact same mutation in the family, because we're all genetically different to some, so there's probably modifiers and other things going on. However, if I can see that in this one family who I know [harbor that specific] mutation [then I know that] if multiple people walk into the clinic and have variations in their family histories and in their personal history of cancer, that we are seeing a wide phenotypic variation.

Now, instead of testing 1 gene at a time, we will test 30 or 50 genes at a time, and you can pick up some of these less penetrant genes that are causing the phenotypic variation. Sometimes there are major penetrant genes in these families.

What other barriers to germline testing need to be addressed?

We're always learning. Every year or so we add a few more genes to our repertoire and then, maybe they get on some of these panels. E3 ubiquitin ligase WWP1 is associated with PTEN hamartoma tumor syndrome, which is not on any panels, but the paper was published in the New England Journal of Medicine. We keep learning as we discover more and more of these genes. The more genes that we find tend to occur in less and less people, based on our current knowledge, but some of these patients present with these rare phenomena.

We're also finding out that some of these mutations aren’t specifically a change in the DNA sequence—there are methylation, or rearrangement, or even a deletion. You have to use other techniques in addition to sequencing to figure those families out or those families will be left in the lurch.

The downside of doing multigene panel testing is that now, if you push for more whole-exome and whole-genome sequencing, we have a lot more variants. One commercial lab got [results] back to me 2 months ago from a patient we had tested 4 years ago. They said, “We finally have enough people [where we could determine that] his variant is not significant.” It was good news. We are now more sure of variants because they now have more families in their database at the commercial lab. Sometimes it takes years to figure it out, unless we have functional analysis for all variants. That’s a big challenge right now.

Where do you hope to see the future of genetic testing head?

In a good way, genetic testing will probably [have a lower] cost and there [will be an] ease of doing it [with] whole-exome and whole-genome sequencing. It will even overtake panel testing over time because the machines are better and faster. The key, though, is having a database that you can go back and forth and analyze. You’re going to need the analytics and tools. What happens with the patient? Do I carry this [information] on a flash drive? Is it in a database I have to have access to?

It's not an easy answer and I'm not sure if the health system that a particular patient goes to is going to store all this information—3 billion base pairs of information—and go back to it each time. Each place is going to have to have the right analytic tools to go back and [retrieve that information]. There are going to be some challenges with that, even though that's the way the technology is going.

The more challenging pieces [are related to] direct-to-consumer (DTC) testing. You don't always know what you're getting on those tests. We can test you for common diseases, such as diabetes and hypertension, but we also test you for BRCA1/2. In reality, very few of the DTC [tests] are doing sequencing or panel testing like we do clinically. Many of them are using single nucleotide polymorphisms (SNPs) that give you a suggestion. Many of these start from ancestry companies,and they recently moved into [testing for] these diseases because people are interested. I don't blame them for doing this, but the information they give might only [include] a fraction of the actual disease variants. If someone finds an SNP in BRCA1/2 or Lynch syndrome, you might need to see a doctor. [Based on your family history or phenotype,] we may have to send a ChIP test to verify [the results].

In some cases, people will test just to be curious, and they think they're going to have something, but there is zero evidence—no personal history and no family history. There are going to be some challenges with the DTC [testing] because we don't always know the type of test they’re getting and the information is not going to be as precise and could present challenges in the clinics. Some people are going to get upset because we're going to say, “No, you don't need testing,” or [patients will ask], “Why does this test say I might have it but your test says I don't?” We have to explain all this and those are going to be challenges.

What else would you like to add regarding the evolution of genetic testing?

There is phenotypic variability in the presentation of many of these syndromes. The standard now is multigenetic panel testing to try to assuage the phenotypic variation; we do pick up [genes in] people who we didn't necessarily think had that disease. I've been surprised too many times, so I'm not surprised anymore. A lot of these inherited conditions have phenotypic variability. If you have any suspicion or your primary care physician has any suspicion, feel free to send [a test] to our clinic because we can investigate that and do testing that's relatively cheap if there's a good cause to investigate that. It may save their life and the lives of their loved ones.