Jeremy Segal, MD, PhD, discusses pathologic considerations in non–small cell lung cancer and the evolving role of molecular testing.
Jeremy Segal, MD, PhD
The future of molecular testing in non—small cell lung cancer (NSCLC) lies in combined large-scale DNA and RNA analysis, said Jeremy Segal, MD, PhD, adding that expanded sequencing can increase the detection of targetable mutations or molecular signatures and can also help avoid misdiagnosis.
“Certain things [can only be detected] if you sequence a larger number of genes,” said Segal. “[For example], you can't do a very good job looking for tumor mutational burden (TMB) or microsatellite instability if you're only looking at a small sliver of the genome. Allowing yourself to look more broadly opens up additional biomarkers.”
In an interview during the 2019 OncLive® State of the Science Summit™ on Lung Cancer, Segal, director of Genomic and Molecular Pathology Unit and associate professor of pathology, Department of Pathology, University of Chicago, discussed pathologic considerations in NSCLC and the evolving role of molecular testing.
OncLive: What is the role of molecular testing in NSCLC? How do you approach it in practice?
Segal: Most of our orders for molecular testing come from oncologists; [these requests are] not typically driven by pathology. However, we have a nice agreement set up with our oncologists at the University of Chicago, where any lung adenocarcinoma that comes through will get automatically reflexed for testing, and then we will [conduct the testing]. We have been seeing many of these cases. We have a nice system that we use where we do an aggressive molecular analysis of all of these tumors, which means that we do a large-scale DNA panel, as well as a large-scale RNA sequencing panel, to look for gene fusions.
[This approach has] allowed us to see a variety of different things by looking more expansively into these tumors. We have repeatedly seen some interesting findings in certain cases. Sometimes [our strategy] is helpful for sorting out a tricky differential diagnosis; it can help us determine if [a patient’s tumor] is a metastatic tumor, a new metastasis, a new primary, or just some sort of spontaneous finding that helps us reconsider the whole case. Is it not lung cancer at all? Is it a metastatic tumor from somewhere else? We have found [ourselves in] several of those [situations] and it has made us—at least me—rethink one of the main purposes of this testing.
We went into designing [these tests] to be biomarker tests for treatment; they are that, and they're good at that, but they're also general pathology tools that we have been able to use in our Department of Pathology. We have an integrated service where we do the pathology, perform the molecular testing, talk about [the results], and compare notes on cases. This has helped us to rethink several cases. We actually ended up changing the diagnosis for a decent subset of these patients; it's probably just 2% [of the overall population of patients with NSCLC that we see], but 2% equates to a lot of patients with cancer.
Could you expand on some molecular testing methods that you use in lung cancer?
We do a large-scale DNA, next-generation sequencing panel where we sequence about 1200 different genes to look for mutations, copy number changes, TMB, and microsatellite instability, among others. Then, we also run an RNA-sequencing assay, in which we sequence the same 1200 genes, but instead we can see if 2 genes were broken and put together in the wrong way to make a fusion RNA transcript. We can find those types of fusions in a large number of genes rather than just a couple. This is also helpful because, every once in a while, there is some random strange thing that you might find in one of these cases.
Why is testing not as common in squamous cell carcinomas as it is in adenocarcinomas? What are the benefits of molecular testing in the former?
Generally, people think of just doing molecular testing for adenocarcinomas because they have the classical biomarkers that people talk about, meaning EGFR, KRAS, ALK, RET, ROS1, and NTRK fusions. Many treatable markers exist in adenocarcinoma; this is not so much the case in squamous cell carcinomas. As such, labs don't typically really test [squamous cell carcinomas] that often.
We haven't tested too many of them, but among the ones that we have tested, we have found a few cases where instead of it being a lung squamous cell carcinoma like we had originally thought, a small subset of patients have the mutational signature of ultraviolet light exposure. This means that they probably came from skin primary. We have seen these hypermutated tumors with this ultraviolet signature a couple of times, where there was an original question about whether it might be a metastatic skin cancer. However, in a couple of cases it was not [suspected], and it was just a random surprising finding. Cases that are skin squamous cell carcinomas tend to respond quite well to immunotherapy.
How do you see some of these testing methods evolving?
Over time, labs will do a larger amount of analysis. Certain labs still do smaller gene panels; we do one of the larger ones. [New options] are coming out; commercial kits are now helping labs do larger amounts of sequencing.
We also have a consortium that we have built of sites that are working to expand the amount of sequencing. Costs are going down as well; it doesn't cost that much to sequence more. The question is, “How much [does it cost]?” I don't know, but moving into the RNA space seems worthwhile. I believe more labs will be doing that; it's a little technically challenging to get into it. This is because RNA from formalin-fixed tissue can be difficult to work with, but better methods are coming out. We will see an emphasis on a combined DNA and RNA analysis; that will be something that more labs will do over the next few years.
What is your take-home message regarding molecular testing?
We’re seeing value in having our in-house testing capabilities and getting the data [from those tests]. Keeping all of that within pathology and considering this testing as a core component of pathology rather than just a biomarker service that anybody can do anywhere [is important to remember]. There are some nice [benefits] to having [this testing] integrated into whatever pathology service you use as an oncologist. By doing that, we are able to compare notes with one another. We can see if there's something wrong with a case, we can learn, and we can teach each other. Pathologists regard genomics and molecular testing as the future of our field, and I would just like to emphasize that point. [Molecular testing] is a pathology tool in addition to an important biomarker testing tool.