Clinicopathologic Features and Response to Therapy of NRG1 Fusion–Driven Lung Cancers - Episode 1

NRG1 Fusion–Driven Lung Cancers: Clinicopathologic Features and Response to Therapy


A review of data from the eNRGy1 global registry study, with deference to the efficacy of current therapies and clinicopathologic features.


Alexander Spira, MD, PhD, FACP: Hello, and welcome to [CancerNetwork®] Between the Lines, a journal club experience. Today’s featured article is “Clinicopathologic Features and Response to Therapy of NRG1 Fusion-Driven Lung Cancers: the eNRGy1 Global Multicenter Registry.” My name is Dr. Alex Spira, a medical oncologist at Virginia Cancer Specialists [in Fairfax, Virginia] and USOncology. Joining me is my colleague Dr Ross Camidge, the director of the thoracic oncology at the University of Colorado [in Boulder, Colorado]. Welcome, Ross.

Ross Camidge, MD, PhD: Hey, Alex. By the way, it’s eNRG. See what a clever word play it is?

Alexander Spira, MD, PhD, FACP: I like that. It was thought of just because of that, right?

Ross Camidge, MD, PhD: Yes, you’ve got to find your acronym that works.

Alexander Spira, MD, PhD, FACP: For today’s featured article, Dr. Alex Drilon from Memorial Sloan Kettering [Cancer Center] was the lead author, and we’re going to be discussing the article. Ross, let’s talk about the clinicopathologic characteristics. This paper looked at a registry in which physicians submitted their patients with these NRG1 fusions. Why don’t we talk a little about the clinical pathologic characteristics?

Ross Camidge, MD, PhD: Essentially, what happened was Alex Drilon and a couple of other people started to email their friends who worked on this other registries. They said, “Have you got any cases of NRG1 fusions?” Then they tried to pull all these rare ones together to see what they look like. You can see that there’s no real sex difference within the sites that we looked at. There was no ethnicity difference that stood out. There’s a preponderance of never smokers. As you might imagine with these driver oncogenes, they tend to be simpler cancers, but it did occur in some people who were smokers. One of the things I thought was interesting is they talk about the median age. Often we think of these driver oncogenes occurring in younger patients. For me, of the 3 or 4 cases I’ve seen, they’ve tended to occur in older patients. They’re more like the MET exon 14 skip mutations. You can have rule breakers that have a smoking history, and you can have rule breakers who tend to be older. They’re mostly adenocarcinomas, as you might imagine, and you can see all the countries that we accrued this registry from.

Alexander Spira, MD, PhD, FACP: Considering 110 patients are accrued, it’s pretty impressive if you think about the incidence of this. You had 3 patients. It must have taken a lot of work to help identify all these patients from everybody trying to remember who they are, right?

Ross Camidge, MD, PhD: What happened is that the diagnostic evolved in recent years. So using an RNA- in addition to a DNA-based extraction to do next-generation sequencing boosts the sensitivity for picking up these rare fusions. We got a bit of a flurry because we backfilled some of the patients who didn’t have a driver oncogene, but we found some. Then our hit rate dropped down to the low level you’re expecting. These are running less than 0.5% of lung cancer if I recall.

Alexander Spira, MD, PhD, FACP: I’ve always been impressed by hearing you and others from your institution speak. You were 1 of the biggest proponents of RNA-based testing before almost anybody. Now it’s more routine at some of the commercial labs, but you were 1 of the proponents. I remember hearing you guys talk, and it persuaded me to make sure when we did our commercial base that we were using RNA-based testing.

Ross Camidge, MD, PhD: People sometimes get confused about it. All you’re doing is changing the starting material for your next-generation-sequencing assay. You take DNA, stick your primers over exons, and hope these things join up. When you use an RNA-based extraction as well, your primers can run into an unknown territory, because you’ve gotten rid of the introns. You can pick up fusions and MET exon 14 skip mutations much more easily. It’s adding to the extraction. You’re not saying, “I’m just going to do RNA,” or “I’m just going to do DNA.” You’re going to say, “I want both as my starting material.”

Alexander Spira, MD, PhD, FACP: That’s great. Thanks for pointing that out. We’re going to look at some of the clinicopathologic features of these patients. The frequency of metastases with patient with NRG1-fusion-positive lung cancers and how it breaks down. My take-home on this is they have some atypical metastases, soft tissue, sinus, and some atypical spots; but all in all, there’s nothing here that stood out to me. What were your thoughts?

Ross Camidge, MD, PhD: By the time you got down to the stage IV patients, there were just over 40 patients. You look at the pattern of metastatic disease at diagnosis so you can remove the concept if you live longer and the disease spreads more to places. This is a pattern of metastatic disease at diagnosis in just over 40 patients. There are retinal lesions. That’s a bit weird, but they’re only 2% of cases. It’s hardly diagnostic of an NRG1 fusion. So I agree, nothing leaps out.

Alexander Spira, MD, PhD, FACP: Then I look at the survival, and the N [number of individuals in the study] is relatively small. I always say that it has to be taken with a grain of salt because these are patients who manage to get themselves to big academic places, so they probably are in a little better shape. It appears that they do about the same. We’ll learn more once we get into the real world and see how these patients do.

Ross Camidge, MD, PhD: You’re absolutely right. For the patients who find their way to an academic center through several other oncologists, with no driver identified, and then turn out to have NRG1, have declared themselves the survivor athletes to get there in the first place.

Alexander Spira, MD, PhD, FACP: In terms of phenotype, there’s still a lot of PDL expression unknown. This database probably goes back for a little while. It’s almost unheard of that you would have PDL unknown given what we’re doing. What are your thoughts here?

Ross Camidge, MD, PhD: When you’re looking for something rare, you’re looking down the back of a sofa or in the back of a cupboard just to find these cases. These are archival specimens—people have pulled it out with testing, possibly after the patient has passed on. That may explain the PD-L1. If you’ve got enough tissue for a genotype, then you should have enough for PD-L1.

Alexander Spira, MD, PhD, FACP: I always have this dichotomy of what we think of as a driver-mutation-driven tumor. There are things. One is the PDL expression, and I’m sure you experienced the same thing. How many L [PD-L1]–positive patients have a high PDL score? That doesn’t necessarily correlate, so I’m not so sure what to make about that. I don’t think there’s anything specific or special about these tumors.

Ross Camidge, MD, PhD: No, but it raises a point. If I have a never smoker with adenocarcinoma who has PD-L1 of a 100%, for me that’s a red flag that they have a driver oncogene. You have to resist the temptation to rush for your single-agent immunotherapy and wait for the molecular to come back if you possibly can.

Alexander Spira, MD, PhD, FACP: That’s a great question, Ross. To this day, I still get those questions, despite all these talks you and I have done and other things. The patient is ALK and has a PDL of 90%. What would you do? You almost have to hold back what you want to say, but make sure you treat the targeted therapy, not the PDL status.

Ross Camidge, MD, PhD: When I speak to pathologists, one of the things we’d like to have is something that tells us if there are tumor-infiltrating lymphocytes there. PD-L1 in the absence of lymphocytes is almost certainly downstream of an oncogene. PD-L1—when there are tons of lymphocytes, there’s almost certainly an immune system primed to attack the cancer that’s been fended off by PD-L1. Having some kind of CD8 stain sounds easy, but the pathologist says that’s hard to quantify. But that’s what we’re missing.

Alexander Spira, MD, PhD, FACP: This is an aside, but we’ve been trying to do better than PD-L1 staining since nivolumab was first approved, and we’re still not there.

Ross Camidge, MD, PhD: Yes.

Alexander Spira, MD, PhD, FACP: Great.

Ross Camidge, MD, PhD: On this slide, on the right hand side, this is tumor mutational burden. Alex Drilon is at Memorial Sloan Kettering, so it was a bit of a product placement for MSK-IMPACT, which includes tumor mutational burden. Only the ones who got on MSK-IMPACT were included in this list. That’s 11 of the NRG1 fusions. They can say it was low, certainly even among the driver oncogenes. That’s not a big surprise.

Alexander Spira, MD, PhD, FACP: That’s my take-home message. Of the 5380 patients, 11 had NRG1. So you can see why we’re struggling. Obviously the ultimate question is, in the absence of an approved drug—in our last journal club, we talked about some clinical studies and new drugs. Now we’re talking about the standard of care, which is systemic therapy. What are your thoughts on how this may or may not differ from some of the other treatments we have?

Ross Camidge, MD, PhD: This is a pretty depressing slide, and we can pull it apart a little. The implication is that these patients, at least in this data set, are not responding to many of the standard therapies. We’ll come back to the afatinib in just a second. One would like to drill down a little. Are these people who have survived long enough to get to an academic center? Do they have incredibly indolent disease, despite not responding? That’s the disconnect. If you didn’t respond, how did you survive to get to 1 of these centers? Then one would want to pull apart particularly the doublet-based chemotherapy. We’ve known for a number of years that some of the fusions, particularly ALK, ROS, and RET, have exaggerated sensitivity to pemetrexed. NRG1 is a fusion, except it’s extracellular as opposed to intracellular, like most of those other fusions. I don’t know which doublet was going on here. I don’t know if they included pemetrexed and, if they did, whether pemetrexed maintenance was involved. There’s a bit of detail. Look at the Ns; they’re pretty small. It doesn’t make me go whoop-de-do. We’ve got lots of great standard therapies. I wonder if it’s as bad as it looks.

Alexander Spira, MD, PhD, FACP: That’s fair. In your experience, you probably have more than I, do you find these to be more indolent than some of the other tumors? What’s your experience?

Ross Camidge, MD, PhD: The ones I’ve had have tended to occur in older patients, often into their late 70s and 80s. It has been fairly indolent, as sometimes we see with cancers that occur in older people. I don’t know why, but that’s not an unusual observation.

Alexander Spira, MD, PhD, FACP: That may be why the response rate is low, but I agree. Many of these columns are single digits, so it makes us think. Then we have some swimmer plots looking at the same thing. Any thoughts? Do you want to talk about the afatinib?

Ross Camidge, MD, PhD: Let’s remind ourselves what NRG1 fusions are supposed to be. NRG1 is a ligand that binds HER3, and the NRG1 fusion essentially tethers the ligand to the surface of the cell. You’ve got this constant supply of a stimulating molecule that’s binding to HER3, and HER3 partners with other members of the HER family and then creates a signaling dimer. Afatinib is a pan-HER inhibitor; sounds great. HER3 doesn’t have an active kinase domain, but the things it partners with—EGFR and particularly HER2—could be inhibited by afatinib. I guess we have proof of principle. We have about a 25% response rate. The numbers are pretty small. It wasn’t 60% or 80%, so is that telling us something about afatinib? Is it telling something about the underlying heterogeneity in this population? Remember, there are multiple NRG1 fusions. I don’t know. Its kind of good news. In theory, this is a targetable oncogene, but it’s the beginning of that journey, not the end.

Alexander Spira, MD, PhD, FACP: In our last presentation that we did, you described it very well. We had a nice schematic of how these work. You have to visualize what you just said in your head, and you did it very well. The afatinib data are interesting. The field hasn’t moved much farther in looking at afatinib, and there are obviously newer drugs. In the systemic therapy activity, there are no surprises. It’s a little less blue than we would like to have seen, but we saw that on a previous slide.

Ross Camidge, MD, PhD: Yes.

Alexander Spira, MD, PhD, FACP: Some conclusions. From what we know, these are relatively rare tumors. There’s not much known. They may be slightly different. Maybe they are slightly slower growing and may not be responding as well. My big take-home is we need a lot more information. The Ns in all these are really small. Some of it may be old registry data. It would be nice to see how these do. As these newer studies develop, we’ll certainly learn a lot more.

Ross Camidge, MD, PhD: The thing that you brought out is that you do this registry, and everybody has an NRG1 fusion and submits it, but you don’t have a denominator. Pulling out what they had in the MSK-IMPACT gives you that clue as to how rare these things are with 11 cases of over 1000 that they had data on.

Transcript Edited for Clarity