Benjamin Levy, MD: Let’s move on to another rare yet actionable mutation, NTRK fusions, an incredible story, the basket study. Enrolling patients I believe 7 months of age all the way to 70 years of age, all harboring NTRK fusions, all treated with larotrectinib. Bob, do you want to talk to us about that data that altered the paradigm and walk us through some of the other drugs that are coming out, including entrectinib for NTRK fusions?

Robert Doebele, MD, PhD: Absolutely. It’s really a family of fusions, so we’re talking about 3 different genes—NTRK 1, 2, and 3 fusions. They’re all relatively rare, we’re talking about a total incidence here of about 0.3%, although it’s 0.3% not of lung cancer but really across all tumor types. In fact, I’m not sure that there’s a tumor type that we haven’t found one of these fusions. That creates a little bit of a testing dilemma for us.

They’re relatively rare, but they follow the same paradigm of ALK and ROS fusions in terms of their structure. It’s also led to some challenges in terms of testing. There are many ways to test for these. I think there’s a lot of disagreement. The first drug that was FDA approved does not actually have a companion diagnostic, probably lending to more confusion about how to test for this. There’s a lot of possibilities. You can screen by IHC [immunohistochemistry], but you really have to follow that up with an NGS [next-generation sequencing] because that is not a very specific test. You may have TRK expression.

Even the sensitivity isn’t fantastic in NTRK3. NGS is the best way to go. RNA-based NGS is probably the best way to go, but not a lot of places are using this. There are some commercial assays at our institution that use it. I think DNA-based NGS is a reasonable approach to this. But especially if you’re wanting to choose a large panel that has hundreds of genes on it and you’re looking for ALK, EGFR and all these other things. I think that NGS is the best way to go.

Larotrectinib was the first drug and it is now FDA approved. It’s actually the first oncogene targeted therapy to get a tumor agnostic indication, of course. Pembrolizumab was the first to get a tumor agnostic indication for MSI [microsatellite instability]-high. Indeed, they enrolled patients from pediatric all the way through adult, and also multiple different tumor types. The only requirement is you needed to have an NTRK fusion, and that really tells us that the biology is the most important factor here. The response rates were in the 70% range. The PFS [progression-free survival] is not yet evaluable in some of these studies. We’re seeing fantastic activity in a unique situation where the oncogene is driving the boat here, it’s positive in this lung cancer, breast cancer, whatever else.

I think there’s a lot of excitement about this, but some confusion about how to test for it. I think in lung it’s easy. We’ve got so many other genes that we need to be testing for. I think NGS makes the most sense, a large panel NGS assay. I think for some tumors where routine testing is not common, it creates a bit more of a dilemma. There’s such a great benefit for the individual patient but a relatively rare driver.

Zofia Piotrowska, MD: Any clear clinical characteristics that point who to test for?

Robert Doebele, MD, PhD: No. I think it’s been challenging in terms of what type of patients to look for. There are a few rare tumor types—mammary analogue secretory carcinoma, NTRK fusions, in particular NTRK3 fusions seem to be pathognomonic for that. If you have something that looks like a secretory breast or a mammary analogue secretory carcinoma of the salivary gland in head and neck, those are likely to be, in fact they’re almost certainly ETV6-NTRK3, and there are a few pediatric cancers also that seem to have pathognomonic. But in some of the larger cancer types where we rarely see these: lung, thyroid, colorectal, it’s not clear that you can pick out the patients who are going to have them.

Benjamin Levy, MD: What about entrectinib?

Robert Doebele, MD, PhD: That drug’s not yet FDA approved, but also seen fantastic activity. Some people have compared the response rates; they are little bit lower, 57% in the entrectinib study, but that was only adults. We know now actually after some more data have come out, in fact, larotrectinib showed adult-only data, and the response rates have come down a smidge. And then entrectinib has presented pediatric data where their response rate was 100%. I think it tells that you really need to think about those populations separately. It’s really hard to cross-trial compare when you’re taking about a dozen different tumor types. It’s even harder when you include pediatric patients.

But the real big differentiating factor with entrectinib, as we discussed with ROS1, because it’s also a ROS1 inhibitor, is the CNS [central nervous system] activity. It has really good activity there. I think we’re still waiting to see that from larotrectinib if there’s intracranial ORR [overall response rate] or PFS that’s meaningful with larotrectinib. I think for patients who are at risk for brain metastases, certainly cancers like lung cancer where there’s a high rate of brain metastasis, I think entrectinib may have some traction in those types of tumors.

Benjamin Levy, MD: And then finally round this out with LOXO-195, which was presented by David Hyman, MD, at AACR [the American Association for Cancer Research annual meeting] in the resistant setting. It’s for a rare genotype but a drug that is highly active.

Robert Doebele, MD, PhD: As we were thinking about developing a targeted therapy, we could look back to EGFR and ALK and know that kinase domain mutations were going to be a problem here. So I think this was a very smart development strategy to have a follow-on drug that has activity against some of the kinase domain mutations that we see, solvent-front mutations. That’s what LOXO-195 was really meant to cover, the patients who have acquired resistance through a TRK mutation in the same way that we see gatekeeper mutations with ALK and EGFR.

And we’ve already seen data. David Hyman presented this at AACR. And we’ve seen activity in the mutations where the response rate was around 40%. I think one thing that was interesting, going back to our discussion about ALK and testing, is that one of the things that they looked at, when there was an identified bypass signaling mechanism going away from TRK to something else, the response rate was 0%. I think that really does tell us that we can use the biology to sequence these drugs, but it’s exciting to have a follow-on option so quickly come in in the pipeline.

Benjamin Levy, MD: Crowed space for a rare genotype. I’m still looking for my NTRK fusion. I haven’t found one yet.

Zofia Piotrowska, MD: I’ve heard it’s one per career.

Benjamin Levy, MD: I’m still going, I’m not defeated.

D. Ross Camidge, MD: Do we have 3?

Robert Doebele, MD, PhD: I think I’ve treated 3 so far.

Benjamin Levy, MD: I think you said this is the fusion that’s predictive, not a mutation. We’ve gotten a couple of these just mutations back, which are more common, which is not predictive of the drug’s activity.

Robert Doebele, MD, PhD: Absolutely. I get asked that question a lot, and you see mutations on the report frequently, and I think it can be very confusing. It’s an NTRK fusion or an NTRK rearrangement that it really needs to be called.

Transcript Edited for Clarity