John L. Marshall, MD: Tony, let’s switch to you for a second. Talk us through the measuring of NTRK fusions. We thought about mutations and amplifications, and now we have fusions. I know for a lot of oncologists, including me, there’s been a learning curve of what this is all about. Walk us through some of that background.
Tanios S. Bekaii-Saab, MD, FACP: When we talk about these alterations that affect the NTRK pathway, they go to this big bucket. A very small proportion of fusions, actually up to 14% of tumors, will have a mutational amplification or alteration. The question is, of course, how important are the nonfusions in terms of promoting oncogenesis? That comes with a question mark.
The fusions seem to be drivers. Not as much is known about the mutations and amplifications other than the fact that, for example, the presence of an NTRK mutation could be a driver for resistance for targeting these NTRK fusions, meaning it could be an escape mechanism.
There are some agents that are more globally attacking at least in early development, both the fusion and the mutation, so pan-targeting that may actually have an advantage, versus those that target specifically the fusion. For now, most of our understanding is linked to the fusions and the roles in oncogenesis and targeting them. Less is known about the others and targeting them.
John L. Marshall, MD: Perhaps now the GI [gastrointestinal] specialists can gather together. Shub, jump in on this. At least in our world we’re doing a lot of MSI [microsatellite instability] testing, and we see whether it’s PD-L1 or MSI. Those are in the colorectal cancer world where we often see the NTRK fusions. I’ve always just thought this is because there’s just a lot of shrapnel around with an MSI setting, and so you’re going to more likely have 1 of these fusions. What’s your take on that, Shub? You think you run them together?
Shubham Pant, MD: Right. The issue, again, is it’s like finding a needle in a haystack for 1 of these NTRK fusions. As you know, I’m in 2 departments at MD Anderson: ICT [investigational cancer therapeutics] and in GI medical oncology. A lot of my patients on NTRK inhibitors have been on the phase 1 side. We’re just introduced them 3 or 4 years back when we were surely seeing the early responses to the NTRK inhibitors. So we always ask this question in different meetings. People always ask us, “It’s such a small percentage, why pay attention to that group?” But where we can really enrich them, you’re exactly right. In the MSI-high population for rectal cancer, they can have a higher percentage, and that’s a relative term of these NTRK fusions.
Small pathology studies, which were done in patients who had NTRK fusions, as much as 75% of them were also MSI high. This MSI-high group really has 2 tumor-agnostic approvals for the immunotherapy and for NTRK fusions, which is very interesting.
John L. Marshall, MD: Then you find them in the same patient. Luis, what’s your take in lung cancer? Are you seeing this type of overlap with some of your molecular targets, or you’ve still got your nice pie chart where everybody is in their own wedge?
Luis E. Raez, MD: No. We are very fortunate that in lung cancer we have several molecular targets. The problem is that for someone who has squamous cell carcinoma, a lot of targets don’t have drugs yet. What I want to add also about this conversation, what Tony was mentioning before, is something very practical for the audience. These NTRK mutations get reported all the time, and sometimes even the amplification. For example, our site was 1 of the original accrual sites for LOXO-101, larotrectinib. It was the only site in South Florida.
A lot of community people just were sending me patients for accrual in the study, but they didn’t notice that NTRK mutations are not NTRK fusions, and it’s not an NTRK amplification. As much as they know even with entrectinib, there are data that have people treated with NTRK mutations and NTRK amplifications, and they don’t have a response. That’s why I think it’s a very important practical point. We only treat NTRK fusions unless it’s a patient with a fusion that has progressed and is not the mutation originated resistance. As you said, as clinicians, we are learning very fast about all these things.
John L. Marshall, MD: That’s such an important point, because lots of us have sent off molecular tests on our patients, say a year or 2 ago, and they’re still sitting scanned into the back of our EMRs [electronic medical records]. We think we might have done this test, for example, when in fact we were looking at either protein or mutation data and not looking at the fusion. As I think somebody else said, unless it’s the fusion, the medicines aren’t really working, so it sometimes justifies retesting.
Tanios S. Bekaii-Saab, MD, FACP: The mutations, again, invoke resistance to these agents, so in many ways the presence of the mutation not only does not predict a response, but it does predict for resistance. It’s interesting. When you look at the tumor mutational burden overall, we always think that these alterations, these fusions are always present or co-expressed with a number of different alterations, whether it was cell cycle as treated, tyrosine kinase, MAP kinase, PI3K. You can find a lot of co-localization. You find PD-L1 expression in about 20% are MSI high, about 20% to 30%. So you find a lot.
But the tumor mutational burden overall is lower for the NTRK-fusion-related or driven tumors, versus those that are not. Although they co-locate, they’re not necessarily loaded with mutations as much as initially thought.
Mark A. Socinski, MD: Yeah, that was my point below. I don’t know that we know as much as we need to know about the NTRK lung population. But if you look at other fusion-like things, they tend to occur in the never-light smokers, almost always adenocarcinoma. They almost always have low PD-L1, they have low infiltrating lymphocytes and tumors, and so forth.
We largely ignore PD-L1 status. If you diagnose a driver, the use of targeted therapy in the driver setting, it’s so much more effective than what we understand immunotherapy therapy to be in this setting. So we largely ignore PD-L1 in that population, but it’s so infrequent in lung cancer, I still think we’re learning.
John L. Marshall, MD: Yeah, I think that’s true too.
Transcript Edited for Clarity