Case: A man in his mid-50s with metastatic non-small cell lung cancer (NSCLC) and an ALK rearrangement receives brigatinib as first-line therapy on the ALTA-1L trial

DR LOVE: I met with Dr Ross Camidge, and to begin, I asked him about a patient from his practice.

DR CAMIDGE: The first guy I’ve actually known almost as long as I’ve been in Colorado. He’s now 55, and I first met him back in 2006, where he thought he’d had pneumonia, and then they had the usual workup. And he was what appeared to be a Stage I cancer, which was resected. It was a 3.6-cm invasive cancer, back in the days when we used to call it bronchoalveolar carcinoma, node-negative. That was resected January 2006. Because it was node-negative, it was still — it was 3.6 centimeters. He had no additional adjuvant therapy. And we just followed him. And actually followed him for 8 years, and then he got bored of me, and then he just disappeared from follow-up.

DR LOVE: Could I just ask 1 thing? I noticed here in his history that’s kind of interesting — so he was a smoker, but he was still smoking?

DR CAMIDGE: Yes, no, I mean, he’s a great guy. He’s a rough diamond. He’s definitely kind of like, “Yes, doc, I know I should give up the smoking,” and then he’s got the packet of cigarettes in his top pocket. He tried and failed to quit multiple times.

DR LOVE: Wow.

DR CAMIDGE: He’s a nice guy, and we reached a detente in terms of me trying to nag him too much on that.

DR LOVE: You just kind of gave up on it after a while, huh? That’s really amazing, though.

DR CAMIDGE: It’s really interesting, because then what happened is something like 12 or 13 years after he’s first diagnosed, he said, “I just felt I should go back and do a follow-up.” What happened? He says, “I was moving this refrigerator and felt I pulled something.” And of course he comes in, he has a chest scan, and now suddenly there’s a right-sided pleural effusion. There are lymph nodes in the middle of his chest, and there are nodules in his right lung.

We did a PET scan, and there’s stuff everywhere, including with peritoneal spread. The lymph nodes all light up. These nodules light up. And he gets a biopsy, and I’m thinking, gosh, this has to be a second cancer, 12, 13 years later. And not only is it the same cancer, but we prove it by doing molecular testing. He has an ALK gene rearrangement, a rare one. He has something called a HIP1-ALK fusion rather than the traditional EML4. We went back to the 2006 specimen — exactly the same gene rearrangement. Interestingly enough, that diagnosis of Stage IV cancer — he quit smoking in a heartbeat.

DR LOVE: Wow.

DR CAMIDGE: He just needed the right motivation.

DR LOVE: Wow.

DR CAMIDGE: Brain scan was clear. And so we’re in this situation where suddenly — this was happening in about October 2019.

Results from the Phase III ALTA-1L and ALEX studies evaluating brigatinib or alectinib versus crizotinib for patients with advanced NSCLC and an ALK rearrangement

DR CAMIDGE: We had these 2 competing data sets. We had ALEX, which I was involved with. Obviously alectinib had become the standard thing you would go for. But we were in between the 2 interim analyses for the ALTA-1L study, which was the brigatinib versus crizotinib study.

And so I could sit down with this guy, who I’d got to know over nearly a decade, and say, “Look, here’s the data. Which one are we going to start you on?” And at that point in time, which was before we saw the second interim analysis, I said “Look, the data so far, these things look about equivalent. One is maybe a little bit better tolerated than the other. One is 1 pill a day. The other is 8 pills a day. And he kind of looked at me with that okay, why are you even telling me this? I’m going to have 1 pill a day. I’m going to have the one that’s slightly better tolerated, and then we’ll forget about the science. We’ll do the rest later.

He started on brigatinib, tolerated it fine. As I said, he quit smoking. And he’s responding on his first scan.

DR LOVE: No tolerability issues at all?

DR CAMIDGE: No, none.

DR LOVE: Nothing whatsoever. I guess the first question is about the biology of this to start with. He’s a smoker, but he has ALK, and he has a recurrence all these years later. Have you ever seen anything like this, and any thoughts about it?

DR CAMIDGE: The smoker, yes. I mean, I don’t buy into the thing that you should only test never smokers. If you have non-small cell lung cancer, I’m going to test you for a molecular driver, because we’re constantly surprised. The hit rate is absolutely less, but I’ve had people who absolutely do not fit that textbook definition. They’re older. They have comorbidities. They’re heavy smokers. And this guy, he’s a current smoker, and yet they’re ALK-positive. It’s surprising, but it’s not undescribed.

In terms of the biology, yes, I don’t think I’ve ever seen a relapse that happens more than a decade after and it’s still the same cancer. He must just have a relatively slow-moving cancer.

DR LOVE: Yes, what a really amazing case. And I guess the other thing here that you were alluding to is the choice that you have. Can you talk about right now what data we have in the first-line setting for patients with ALK disease?

DR CAMIDGE: I’ve been giving this quite a lot of thought over the last couple of months. In general the first thing you should say is when you actually try and compare like with like, the similarity between the first-line efficacy data for alectinib and brigatinib is incredibly similar. I mean, it’s almost freakishly similar. Yes, you can play the game by saying, “Look at my investigator-assessed endpoint. Look at my independent radiology review-assessed endpoint.”

But when you try and compare like with like, here is the data: The IRC-assessed hazard ratio for alectinib when it was first done was 0.5. It was 0.49 for brigatinib. If you look at the investigator-assessed hazard ratio at the updated data analysis, it was 0.43 for alectinib. It was 0.43 for brigatinib. If you look at the 2-year progression-free rate by the investigator, it was 56% for alectinib. It was 56% for brigatinib. And the only thing that there’s a subtle difference is the point estimate of the median progression-free survival by the investigator. It was something like 34 months with alectinib, and it was something like 29 months with brigatinib, but the confidence intervals overlap. And there’s no doubt that that one of the data sets that all look identical, the one where they — look, different people used to say, “These things are different.” But there is nothing really to say, “These things are different in terms of the efficacy in the first-line setting.”

In terms of choices, I think you could approach it in a couple of different ways, and you have to kind of decide what kind of physician you are. You almost have to project into the future — if you are going to give your patient lorlatinib in the second-line setting, which is what the NCCN Guidelines say — it’s almost what the FDA indication is — then you literally have a head-to-head choice between alectinib and brigatinib.

Side effects, dosing and quality of life with alectinib and brigatinib

DR CAMIDGE: The only differences are when you look at quality of life: Alectinib was unable to show a statistically significant difference over crizotinib, but brigatinib could, and the other is the pill burden. Now, I’ve talked to patients, and they go, “Pill burden doesn’t matter. I’m talking about a life-threatening disease. Why do I care if I take 100 pills a day or 1?” And you go, “That’s fine. If I said the only way you can stay alive is 100 pills a day, you’ll take it. But if I offered you a choice, most people would probably take 1 pill a day.”

DR LOVE: Alectinib is 8 pills a day, brigatinib is 1.

DR CAMIDGE: Yep. And the tolerability, compared to crizotinib, the only one that’s actually shown a statistically significant improvement is brigatinib. Alectinib looked the same. And that’s certainly my clinical experience. Brigatinib is a somewhat better-tolerated drug than alectinib. These are all pretty good tolerated drugs, though.

DR LOVE: I was going to ask you about your clinical experience, because initially I had this thought that alectinib was no problem at all, like, zero problems, and then I started to hear stuff. I don’t know, edema or whatever, but what do you actually see?

DR CAMIDGE: Some people, I mean, there are people who will have the worst possible drug, and they have no side effects. And there are people who you give the drug that everyone else tolerates, and they have terrible side effects. But in general, most people on alectinib are still aware that they’re taking a medication. They get peripheral edema. They often get exaggerated skin photosensitivity, which varies depending on where you are in the world and what season you’re in. Brigatinib, I would say, occasionally people can get some achiness associated with it. You can certainly get myalgia with alectinib, probably a little bit more exaggerated. And then, I mean, I would say in general if our goal is perfect control of disease and normal quality of life, brigatinib gets closer to that in most people than alectinib does.

Early onset pulmonary events associated with brigatinib

DR LOVE: The one other thing that sticks out there that we’ve been hearing about are these so-called pulmonary reactions. And I see that this man, it looks like he was involved in a study looking at that?

DR CAMIDGE: Yes. I mean, it’s interesting. When we actually — these have been described. They’re running about something like 5% of people get any kind of respiratory symptoms with brigatinib. About 3% get something severe. Unless you’re treating 30 patients a year, you’re unlikely to see this. It’s very specific. It’s literally happening on the second or third day that you go on it. And the amazing thing is, this is not the pneumonitis that we associate with other TKIs. If you go on it and you get it and you stay on the drug, if you can tolerate staying on the drug, within 4 or 5 days they will all disappear.

There’s a completely different pathophysiology. And we’ve done this study. We haven’t presented it publicly yet, but what I can tell you — because this guy contributed to it — is, if you actually do pulmonary function tests, everybody’s DLCO, their ability to transfer oxygen, drops very briefly on day 2, and then it normalizes.

And so I think the people who are getting the symptoms are the people who don’t tolerate that 10% or 20% drop in DLCO for a few days. And that’s what you do. We didn’t know to whom it was going to happen, but if Mrs Jones is a respiratory cripple when she walks through the door, you could say brigatinib may not be the drug for you. Or, if you do do it, you do a shallower step-up, 30, 60, 90 and then 180, and they tolerate it just fine.

DR LOVE: And at this point, what’s your hypothesis about the pathophysiology?

DR CAMIDGE: We don’t know. We’ve got a nice collaboration going on with Cambridge University in the UK, and they’re starting to analyze these CT scans that we’re doing on day 2 in the study, and they think there’s an increase in vascular density. They’re wondering if there’s some vasospastic thing, which then you’d rapidly tolerize to. And brigatinib is one of the drugs that does have some hypertension associated with it, so maybe it has some vascular effects.

DR LOVE: Interesting. And clinically, what do patients actually say?

DR CAMIDGE: Oh, I mean, so if you’re in the severe, you literally feel like your chest is incredibly tight. But probably half of them have mild to moderate, and you literally have to drag the information out of them. It’s just like, I felt a little tight, and then I just ignored it and it went away.

DR LOVE: It sounds like it’s not a big concern from your point of view.

DR CAMIDGE: No. Not at all. I mean, the scariest thing is a severe side effect that’s unpredictable. I think this is a side effect which is 100% predictable, because it happens in everybody, and you can predict who can tolerate it. And then when you think about the pathophysiology, you just want to reassure the patient: If they can tough it out for 4 or 5 days, it will just go away.

Activity of lorlatinib as second- or third-line therapy for NSCLC with an ALK rearrangement

DR LOVE: You mentioned lorlatinib, also. Can you talk about what we know about that? And where we are today in terms of second- and third-line therapy for these patients?

DR CAMIDGE: Yes, I’m glad you brought that up, because I said, “If you decide you’re going to be a second-line lorlatinib person, then the ALTA-1L and the ALEX data is, like, okay, so now we get to choose based on some of these more minor issues, given the efficacy appears the same.”

But there is the concept of what other lines of therapy you could use. Lorlatinib, given post a next-generation inhibitor, was originally reported as having a 40% to 50% response rate and a progression-free survival of about 5 or 6 months. There’s a couple of things to pull out from that. The first is, we’ve recently started to pull out the data, which we have been asking for a while, is to present intracranial efficacy separately from extracranial efficacy. We know that one of the real strengths of lorlatinib is, it’s amazing at getting into the brain. And when you do that, the response rates if you have intracranial disease are probably running 50% to 60%. And if you only have extracranial disease, the response rates are running in the kind of middle 30%.

DR LOVE: That sounds kind of weird.

DR CAMIDGE: No, because what’s happening in the brain is mostly a drug penetration issue. And if you have a drug that’s really good at getting in there, your response rate acts as if it’s an earlier line of therapy.

DR LOVE: But why would it not work systemically?

DR CAMIDGE: Because there it’s been exposed to the drug, so you’ve selected out a biological change in the cancer.

DR LOVE: Wow, interesting.

DR CAMIDGE: I mean, we may go and talk about with some of the later cases, but this whole idea about breaking the body down into intracranial and extracranial compartments is absolutely coming home to roost. The other thing, Neil, is, on the lorlatinib data there’s very nice data that Alice Shaw had led looking at your chance of responding to lorlatinib post a next-generation inhibitor depending on whether you could find a mutation in the ALK, a resistance mutation. If you did, response rate was 60%. And if you didn’t, it was 20%.

DR LOVE: Hmm.

DR CAMIDGE: The concept of one ALK inhibitor followed by another one followed by another one, certainly you can just play the numbers, but the hope is, in the future maybe we’re going to be saying, “Okay, if you’ve got a mutation and we can find it, we feel confident this is the right approach. If we don’t, maybe you’re still going to try it and you’re just going to be a bit more skeptical.” But maybe we’re going to find another group of things that we can do in the people who don’t have a mutation. We will have a choice other than just going to lorlatinib.

Targeting the MET pathway for patients with NSCLC who experience disease progression on EGFR and ALK inhibitors

DR CAMIDGE: We’ve now seen examples of MET amplification, or MET activation, as an actionable second driver post a next-generation ALK inhibitor. I think the field is going to start to fragment.

DR LOVE: I think MET also has been seen in people progressing with EGFR, for example, on osimertinib. Do these patients respond to anti-MET therapy?

DR CAMIDGE: They do. Yes, MET is going to probably be up for the best supporting actor as a secondary resistance driver. We’ve seen it in EGFR, and now we’ve seen it in ALK. And I think one of the things is, as we’ve moved away from crizotinib, which is also a MET inhibitor, we exposed that as a potential mechanism of resistance. What you do is, you stay on your next-generation ALK inhibitor, and you add in crizotinib. Although the hope in the future is, maybe we’ll be using some of these next-generation MET inhibitors, which, for example, also get into the brain, like tepotinib or capmatinib.

DR LOVE: Wow. Have you tried that?

DR CAMIDGE: Yep. Yes, we had a case at World Lung in 2019, really interesting case. Young woman in her 30s diagnosed with ALK-positive lung cancer when she’s in her third trimester. Delivers the baby. We put her on alectinib. She has a great response for 3 months and then rapidly progresses. We try the brigatinib, then lorlatinib. Nothing’s working. Clearly she’s declaring herself as a second driver. It’s not that she has an ALK resistance mutation. We couldn’t find one. We looked for MET amplification, and it was negative. We looked for MET exon 14 skip mutations — negative. But we were able to grow the cell lines, and she had huge levels of MET signaling through a currently unrecognized method. But that was enough to say MET is important to that cancer, and we added in crizotinib, and she’s now about a year and has had a complete response, adding the MET into, in this case, brigatinib for her.

DR LOVE: And you kept the alectinib going?

DR CAMIDGE: Brigatinib, but yes, I kept it going because it was controlling some of her disease.

DR LOVE: Wow. Hmm. She’s on both?

DR CAMIDGE: Yes, she is. Tolerating it fine.

DR LOVE: Wow. What a story. I’m not even going to ask you about this other mechanism, but just to try to keep it under control here, just to get more clinical, first of all, isn’t there first-line data also with ceritinib?

DR CAMIDGE: Oh, it is, but, I mean, I wouldn’t waste a lot of time or effort on that. I think you get these drugs which are kind of proof of principle. Look, going onto a next-generation inhibitor was better than chemotherapy. Historically it looks better than crizotinib, but nowhere near as good as either alectinib or brigatinib. And also less well tolerated. I mean, it was less well tolerated than chemotherapy, so ceritinib has a role in history, but that chapter was pretty short.

DR LOVE: Sure. Right.

Treatment for patients with metastatic NSCLC and an ALK rearrangement who experience disease progression on an ALK inhibitor

DR LOVE: Just to get back to second-line therapy, so it sounds like nowadays you could encounter a patient who starts out with first-line alectinib, first-line brigatinib and probably some people first-line crizotinib that are coming in with relapse. How do you approach these three situations? And for a community-based oncologist, do you recommend any type of repeat mutation testing?

DR CAMIDGE: Postcrizotinib, people go, “Nobody’s starting on crizotinib.” But there are people who’ve been on crizotinib for years and they’re doing great. If they progress, it’s very interesting. Even though alectinib and brigatinib look almost identical in the first-line setting, in the postcrizotinib setting they’re dramatically different. Brigatinib has got nearly double the median progression-free survival as alectinib.

Presumably you select out a mechanism of resistance on crizotinib that only brigatinib has the additional activity against. And so if you’re progressing on crizotinib, probably brigatinib is the drug to go onto, and that’s certainly within its license.

If you progress on brigatinib, I think I would personally try and rebiopsy, and I think one of the problems is, we’re sending next-generation sequencing panels. We’re missing MET amplification in much of that. If you just want to shoot from the hip, try lorlatinib, but if they’re not responding, it’s not just, oh, I’m going to go onto another ALK inhibitor. Start to think that they might have a next-generation hole that you’re missing. Start to think that maybe they have a second driver and you probably have to touch base with an expert center there.

This was the other thing that I wanted to touch on: If you start on alectinib, even though the drug that’s licensed is lorlatinib, there’s now data that brigatinib in that setting actually has some activity. There was a very small paper that came out that I think made people a little worried. There was about 23 patients and it only had a 17% response rate. But we’ve now got data in probably more than 200 patients, and the response rate with brigatinib post a next-generation inhibitor is running 50% to 70%, with a PFS of 5 to 7 months, so very similar to lorlatinib but better tolerated. I think when I first saw the data from ALTA-1L, my first thought was, what I might do in my practice is start someone on alectinib, then do brigatinib, then do lorlatinib. But I’m in an academic center. I’d be rebiopsying and looking for mechanisms there.

Perspective on the potential identification of patients who might experience rapid disease progression on ALK inhibitors

DR LOVE: You also, in your write-up, talked about patients, or trying to identify patients up front who maybe aren’t going to do too well, so-called “bad ALK.” Like, what’s that?

DR CAMIDGE: Where is the research field going to go? We’ve talked about the clinical implications of ALEX and ALTA-1L and lorlatinib and everything else. But what research questions arise? One of the things is starting to realize that you can’t just play the numbers game. Brigatinib postcrizotinib looks better than alectinib but in the first line looks the same. What is going on? Probably if you’re going in before a mechanism of resistance is established, the cancer is much more pluripotential. It can just flip to some other mechanism. And so I think our hope that constantly moving these things into the first line would extend our benefit may be a little disappointing, that we’re seeing.

I think when you look at the curves, though, the other thing we’re starting to realize is that maybe there’s 2 broad populations of ALK. There is a group who is in a plateau of nonprogression. They are about where the median occurs, which is why that number jumps around. And they are going years. They are, if you like, the “good ALKs.” And they have an unmet need, but we’re kicking that can way down the road.

The immediate problem is the ones who’re progressing in the 12- to 16-months’ range, before that median. And what is going on with them? Can we look at something up front that says you’re more likely to be in the kind of bad ALK category, or “BALK,” as one of my colleagues calls it? Maybe it’s going to be P53. Maybe it’s going to be a variant. Maybe it’s going to be something else. But if we could identify those risk factors up front, our potential to do clinical research in the first-line setting will be accelerated, because our median in that group is going to be much shorter. Otherwise, we’d have studies that are going to look like breast cancer, and they’re going to be out decades before we get the result.

DR LOVE: I’m curious at this point, now that we’ve seen these data roll out over the last few years, when you look at the agents — specifically alectinib and brigatinib — that have come out and beat crizotinib up front, how much of that is CNS and how much of it is being a better ALK inhibitor?

DR CAMIDGE: I think it’s a combination of the two. There’s no doubt that probably the reason why alectinib and brigatinib are so much better than the ceritinib in the first line, a lot of that has to do with CNS activity. And the supportive data is, if you look in that first-line ceritinib study and you take the group that had no brain metastasis, the median PFS of ceritinib is actually 2 years. It looks very similar to what we’re seeing in these other studies. And its major Achilles’ heel is progression in the brain, even though it has some activity there.

I think the idea of you have some degree of early clonal equipoise such that you suppress one of these mechanisms and something else pops up — it’s not true equipoise. You do delay it. But I think that has shortened the degree of benefit that we thought we were going to get. I think absolutely CNS is very important, and that has contributed a lot to the benefit of these drugs. And then some of it is trying to suppress mechanisms of resistance. But if you can’t suppress 100%, it’s always going to find a way through.

Role of immune checkpoint inhibitors for patients with metastatic NSCLC and targetable mutations

DR LOVE: I wanted to mainly focus on targeted therapy in our discussion, but related to that, and maybe related to your patient, is the issue of immunotherapy in patients who have driver mutations. And I know you were part of this immunotarget registry, but I’m just kind of curious in general where we are today in understanding the role of checkpoint inhibitors in these patients, checkpoint inhibitors plus chemotherapy and also with your patient here — history, lot of smoking. Do you view that patient differently than you would somebody else?

DR CAMIDGE: I mean, some of these questions we don’t have answers to. What we do know is that the degree of benefit from immunotherapy doesn’t appear uniform across lung cancer. And some of the clues of the group which would derive less benefit and sometimes even no benefit are whether they have a driver oncogene. Now, that’s a confounded variable, because patients with a driver oncogene often tend to have a lower tumor mutation burden. They often tend to be never smokers. We can’t quite pull it apart.

I would say the gene rearrangement diseases — ALK, ROS1, RET — are the best group for seeing the least amount of benefit from immunotherapy. Some of them are explicitly excluded from some of our classical first-line studies with immunotherapy. For that reason, some of them just weren’t analyzed. In the immunotarget, which was an international study, relatively small numbers, but from a monotherapy with a PD-1 or PD-L1 agent, there was a 0% response rate in ALK.

We’re actually in the process of taking the small numbers of people who responded in RET and ROS1 and drilling down on them in detail, because the question you want to ask is, look, is there really a 10% response rate in ROS1? Or is the 1 patient out of your 10 who responded, were they a false-positive on the ROS1 testing? Were they a heavy smoker, as you said? Did they not respond to the ROS1 inhibitor before they went on the immunotherapy? And that’s what we’re in the process of doing.

DR LOVE: But you said something that I don’t believe I’ve heard before, which is separating out the fusions versus the, what, activating mutations in terms of immunotherapy?

DR CAMIDGE: Yes. I mean, EGFR is a good example. EGFR mutations have, in general, less benefit from immunotherapy, but it’s not zero. There are clearly EGFR-mutant patients who’ve responded. And again, if you look at the data set, the hazard ratio in whatever you were comparing it to was always less with the EGFR mutants. Sometimes it wasn’t statistically significant. But it wasn’t on the other side of the line, usually, in the forest plot. Usually it still was deriving some degree of benefit. That doesn’t seem to be the case with the gene rearrangements. It’s still somewhat a work in progress.

The other example is, KRAS is an oncogene-driven disease in many patients. They’re deriving benefit. BRAF is an oncogene-driven disease. They’re deriving benefit. And MET exon 14 is kind of somewhere in the middle, again, because it’s a heterogeneous population. Some are never smokers. Some are heavy smokers.

PD-L1 expression as a predictive biomarker for patients with actionable alterations; approach to the use of immunotherapy

DR LOVE: That’s really interesting. I remember in a previous conversation you referred to patients who had very, very high PD-L1 levels with ALK. Any more follow-up on that?

DR CAMIDGE: Yes. PD-L1, this is the problem, is that when we’re using PD-L1 as a predictive biomarker, we’re making assumptions. We’re saying PD-L1 is up because it’s induced to go up by gamma interferon secreted by activated T cells who are sniffing around the cancer trying to kill it. However, in many of the driver oncogenes, PD-L1 is up in the absence of infiltrating T lymphocytes. It’s up for some reason downstream of the oncogene, and there its predictive potential is zero because it’s not telling you that there are T lymphocytes there.

And I think one of the things that might improve the predictive potential of PD-L1 is to combine it with some very basic immunohistochemistry, and so are there lymphocytes there? And that’s coming.

DR LOVE: For practical purposes, at this time, with not all the information we’d like to have, which is the situation we’re in, it seems like a lot, I am curious when you start thinking about, if at all, checkpoint inhibitors in patients with targetable lesions. For example, in ALK, do you use them at all? Do you use it with chemo? How do you approach it?

DR CAMIDGE: I would say, with the exception of the driver oncogenes, which are traditionally associated with a heavy smoking population, BRAF and KRAS being the obvious examples, in the other ones I tend to deprioritize immunotherapy to about the level of a Hail Mary pass. And the issue when you’re thinking about going onto chemotherapy as to whether you do chemoimmuno, or just the chemo, actually relates to whether you plan to keep the TKI going. Now, you go, “That’s crazy, didn’t we already have that study that you didn’t need to keep the TKI going when you went onto chemo for EGFR mutant?” And we did — was the IMPRESS study, but one of the reasons why we might want to keep it going now is these drugs are, as we’ve mentioned, better at getting into the brain. And maybe what they’re doing is preserving CNS control.

A few small studies came out at ASCO. We have our own data — that actually keeping the TKI going prolongs the PFS when you go onto chemotherapy. And maybe it’s mostly because you’re preventing progression in the brain. And if that’s the case, TKIs and immunotherapy don’t combine well, so that’s the other reason why I’m tending to add in chemotherapy and keep the TKI going rather than stopping it and going onto chemoimmunotherapy.

DR LOVE: Hmm. Let me think that one through a little bit.

Treatment for patients with metastatic NSCLC and EGFR mutations who have experienced disease progression after targeted therapy

DR LOVE: I guess one of the things that you hear about — you mention the differentiation between, for example, EGFR versus ALK. What do you do when you decide that the patient who’s got an EGFR activating mutation’s no longer responsive to targeted therapy? First of all, do you even look at the PD-L1 level in making your decision?

DR CAMIDGE: No. No, I don’t.

DR LOVE: And you hear people talking about using the IMpower150 combination of atezo, paclitaxel, carbo and bev.

DR CAMIDGE: The IMpower150 is the great example of how we’re all flawed human beings, and you’d be surprised how many people — they want to give the IMpower150 regime, but they don’t want to give a taxane. They give carbo/pem/bev/pembro, and they sort of jerry-rig something. You go from a somewhat suspect small subgroup analysis that you could pull apart, if you look at the paper, to a completely data-free zone. And we just do it because we don’t want to give a taxane.

DR LOVE: The issue of whether to use a taxane as part of first-line treatment of metastatic disease for nonsquamous also comes up with the IMpower130 trial of carbo, nab paclitaxel and atezo, which, like the 150 regimen, is approved first line by the FDA for patients without EGFR or ALK genomic alterations with nonsquamous cancer. Potentially you have the option of either paclitaxel or nab paclitaxel.

But let’s get back to the patient with an EGFR activating mutation progressing on osimertinib, no new mutations and you are moving away from targeted therapy. What are you usually thinking about?

DR CAMIDGE: Let’s say they were progressing through targeted therapy. What I would tend to do is keep the osimertinib going, and I would add in carboplatin and pemetrexed or carboplatin, pemetrexed and bevacizumab, and I would keep the TKI going. And I realize I can’t stand up in court and say I can defend this action, but you asked me what I do. That’s what I do.

DR LOVE: And again, even with the EGFR, you’re holding off PD-1 for a Hail Mary?

DR CAMIDGE: Exactly. I think if I did that and they’re progressing, literally I say, “Look, your chance of benefit is about your chance of severe side effects, and so it really is like rolling the dice.” And it’s not quite kill or cure, but it is in the Hail Mary thing. And I have to be honest, most of the time it doesn’t work and you don’t get terrible side effects, and they don’t respond, either. There’s a big middle ground. But I have had people who’ve had terrible side effects, and I personally have not yet had anyone respond. But I have a number of colleagues who have.

DR LOVE: And getting back to this idea of keeping the TKI going. Another thing we’ve been hearing about — I wasn’t aware of until fairly recently — is the issue of, I guess, delayed TKI toxicity in patients who’re getting immune therapy. What about that?

DR CAMIDGE: Maybe one of the things you’re talking about is Mrs Jones walks through the door, you haven’t got the molecular testing back and you’re starting her on carbo/pem/pembro. And then you suddenly find out that they’ve got an EGFR mutation, and you say, “I’m going to stop that. I’m going to stick you on osimertinib,” and they start to get problems because the immunotherapy is still in their system.

And one of my colleagues has described that of poisoning the well, and there’s certainly a move amongst some of us in the field to say, “Look, if you don’t have that information back, and particularly if the patient is a never smoker and you don’t really know or you’re suspicious, don’t include the immunotherapy in cycle 1. Just give them chemo, wait for the results to come back. No driver oncogene, add in the pembro in cycle 2.”

Case: A woman in her early 50s with metastatic NSCLC and an EGFR exon 20 insertion experiences disease progression after treatment with TAK-788 and whole-brain radiation therapy

DR LOVE: Let’s move on now. I’ve always been interested in the exon 20 insertions and hoping that at some point we’re going to be able to do something about it. And it sounds like we’re getting there. How about your 53-year-old lady?

DR CAMIDGE: Delightful lady, 53-year-old Korean American. She’d met her husband when he was over there in the military. First diagnosed in about July 2017, presenting with the usual nonresolving cough, went through the usual workup. But eventually they get imaging, and she has bilateral nodules, liver lesions, a dominant mass about 5 cm in the right lower lobe. She gets a PET scan that shows she has bone lesions as well, a brain scan and she has multiple brain lesions, probably 20 or so at a guess. She has a CT-guided biopsy of the lung. TTF1-positive adenocarcinoma, but with an exon 20 insertion. Interestingly enough, since we’re talking about it, PD-L1 of 70%, but I discount that. She’s a never smoker. She has a driver oncogene.

There are multiple different exon 20 insertions. Hers was a 766 to 769 duplication, though the nomenclature is hard to get your head ‘round. I think if this becomes relevant we’ll learn it, but at the moment we’re just lumping them all together. But there are subtle differences.

Initially goes on carboplatin and pemetrexed and has SRS to 18 lesions in her brain, and she does okay, goes from about August 2017 to about April 2018 with a nice response, so a good response that lasts for a long time until she had some progression in her lungs. At that point in time we have the trial of TAK-788, which is one of these drugs being developed for exon 20 insertions. And so she goes on that in May 2018. The snowstorm of nodules in her lungs, which had been emerging from the mist, dramatically shrinks down. Her breathing gets much better. But pretty much the first scan, so she started in May, and in June suddenly she has new lesions within her brain. And she is logged as progression on the study. She has a very short median progression-free survival per the study.

We’re still desperately trying to avoid whole-brain radiotherapy in these patients who we think might live a long time. She gets 39 additional lesions treated with stereotactic radiosurgery. Her body scan, her extracranial response, is maintained. She stays on the drug in the study until now somewhere between June and December, so about 6 months later, 15 new lesions are turning up in her brain. Again, we treat those with stereotactic radiosurgery. But she is starting to cognitively decline.

Perspective on dementia associated with whole-brain radiation therapy

DR CAMIDGE: Whole-brain radiotherapy treats everything, but if you treat enough lesions by stereotactic radiosurgery, there is some impact. You slow the computer down, and she was starting to slow. We put her on methylphenidate, which perked her up a little bit. And then, unfortunately, she was starting to progress in both the body and the brain. She had symptoms in the brain. She had bad headaches. And we had a very full and frank discussion between myself, the radiation oncologist, her husband and herself. “Look, if we give you whole-brain radiotherapy, we could produce radiation dementia. You’ve had a lot of lesions treated by SRS. You’ve got some cognitive impairment already, but you are getting symptoms from this disease in the brain.” And she said, “Okay, I’m going to try it.”

And she had a terrible outcome, I have to tell you. Literally within 2 weeks of finishing the whole-brain radiotherapy, she entered a cognitive decline. She withdrew. She didn’t have any symptoms, but she had left us and then died about 4 weeks later. And I spoke to the husband a lot afterwards, and we had this long discussion about was it the right or the wrong thing to do? And this is the way we rationalized it. What he did by supporting her when she went through the whole-brain radiotherapy was to take the burden of her disease onto himself. He became the carer. She didn’t have any symptoms. She wasn’t distressed by this. He was the one distressed. And that helped him deal with the horrible situation. It allowed him to think that he had made the sacrifice even though his wife left him, technically, earlier than she might have done cognitively.

DR LOVE: Could you talk about that a little bit more?

DR CAMIDGE: I mean, so for those who’ve seen it, radiation dementia is one of the saddest things that you ever get to see. I mean, literally people withdraw. They stop eating. They are gone even though they’re physically still there. And that was very tough for the husband to go through, because he suddenly became the complete carer. I mean, he had to lift her in and out of wheelchairs and take her to the bathroom and everything else. And he worried that he’d done the wrong decision to push for the whole-brain radiotherapy rather than letting her suffer with the raised intracranial symptoms that she had.

And if you can go back in the time machine, you always wonder if you’d make the same decision. And the way we allowed him to come to peace with that decision was to say we took the symptoms away from his wife, and we shifted the burden of the disease from her onto him, because he now had to be the carer. And that was okay. He was the guy that said, “If I could make the disease happen to me,” and I said, “That’s kind of what you did.” But it was a very tough situation.

DR LOVE: Hmm. Wow.

Let’s get back to the medical parts of this, and I’ll mention that in the past we’ve gone to the Society of Neuro-oncology to do meetings on glioma. We did one on targeted therapy of brain mets. Can you imagine? And then I did an hour and a quarter interview with Greg Riely about it, so that’s oncology. An hour and a quarter on targeted therapy of brain mets. But anyhow, this case certainly brings up a lot of questions.

My mouth is still dropping from the 18 lesions that had SRS, moving up to 35, et cetera. That’s the first question, which is I thought, you don’t go beyond 4 or maybe 10. What’s your experience?

DR CAMIDGE: There isn’t 1 answer. I mean, we are constantly in the process of pushing the envelope. I think in this case we pushed it too far. I think there are other people who have had 100 lesions treated by SRS and have no side effects, so there was something to do with her biology that her normal brain parenchyma was just suffering more with this. It’s not just to do with the number of lesions.

We have certainly shown, in publications from our center, that you can treat tens of lesions and get good control and preserve cognitive function in most people. We are very much trying to avoid that whole-brain radiotherapy, because people are now living long enough that they might manifest radiation dementia from it.

Guidelines from the Response Assessment in Neuro-Oncology Brain Metastases working group: Clinical trial design for systemic agents for brain metastases

DR CAMIDGE: I think the other exciting thing here is kind of what you were hinting at with Greg’s speaking at the SNO meeting — is the concept of really defining the brain as a battleground properly. Nancy Lin, Patrick Wen and myself are some of the architects of the RANO guidelines, or the RANO brain metastases guidelines. And one of the things that came out of things that came out in the most recent set of guidelines, came out in January 2018, was really, how do you define studies? We’ve done lots of retrospective, kind of cobbled together the data, but really prospectively define it.

And the TAK-788 study, I think, is a great example of that. If you look at the data that they presented at World Lung in 2019, they presented the median progression-free survival by whether you did or didn’t have brain metastases at baseline. And one was 3 months, and the other was 8 months. And presumably this patient contributed to that data set.

This is telling you this drug has an overt liability in the brain. And it is much better to find that out early, because they allowed some untreated asymptomatic brain metastases in because now it means when you do the studies going forward, you should be very strict. You should be, as in those guidelines what’s called scenario A, and you should say you can only come into the study if you have no brain metastasis or your brain metastases are treated, because it’s an overt liability of the drug.

The other side is like the lorlatinib study we talked about, where if your drug has an amazing activity in the brain, what you want to do is capture that data as robustly as you can. That’s called scenario B in the guidelines. And that means you’ve got to have regular brain MRIs. You’ve got to capture the response rate in the brain separately from the response rate in the rest of the body. We’re in a brave new world.

DR LOVE: Really interesting.

Hippocampus sparing during radiation therapy for brain metastases

DR LOVE: Another thing that came out that I thought was interesting at the SNO meeting was the issue of whether you can decrease the morbidity from whole-brain radiation. And we polled the audience and the faculty, and it was very interesting the disparity we saw about the use of hippocampal sparing, and I think it’s called memantine, the drug?

DR CAMIDGE: Yes.

DR LOVE: It was really split. There were people using both of these things, not using them, Memorial wasn’t and whatever, but what’s your take on that?

DR CAMIDGE: I think if you have to give whole-brain radiotherapy, most people would try and spare the hippocampus, because that’s where most of your memory sits. To be honest, it’s not absolute sparing. It’s relative sparing. These beams are all going everywhere. There is some suggestion that it has a somewhat better preservation of brain. One of the studies that I referenced earlier, where we looked at people who’d had multiple doses of SRS, we actually calculated the dose to the hippocampus if you’d had 20 lesions treated by SRS versus whole brain versus hippocampal-sparing whole brain.

The dose to the hippocampus is way lower if you have stereotactic radiosurgery, because hippocampal sparing is — it’s a relative sparing. It’s really not completely doing it. The memantine is really interesting, and I thought, when I saw your results, the way it was mostly split was whether you’re a radiation oncologist or a medical oncologist. It seems like the radiation oncologists kind of have bought into memantine, and the medical oncologists are like, “Nah.”

DR LOVE: Yes. Interesting.

Targeting EGFR exon 20 insertions with TAK-788 and poziotinib

DR LOVE: Getting back to these exon 20 insertions, can you talk a little bit more about the biology? My understanding is these are activating mutations just like exon 19, et cetera, it’s just we haven’t had drugs for them. Is that the case?

DR CAMIDGE: Yes. That’s exactly — the biggest challenge is structurally they look very, very similar to the wild-type form of the EGFR. The biggest challenge is being able to inhibit them in a tolerable way. If you remember the story with afatinib that in the lab it could inhibit T790M, but you could never get up to the doses that would do it in a person, because the inhibition of the EGFR wild type, which occurred at a much lower concentration, said you could never get there.

For exon 20, it’s the same problem, but they’re very, very close. I mean, so with afatinib you had the situation that you were having to go through EGFR wild type to get to T790M. With exon 20 and some of these drugs they’re about an equipoise, so you’re always going to get some side effects. What’s happening with TAK-788 is, you don’t have to have more side effects than you have to. With poziotinib, if you like, the proof-of-principle drug, it still has more activity against EGFR wild type than it did against exon 20. It tended to have more in the way of side effects.

Poziotinib’s dose reduction rate at its recommended dose is about more than 60%. TAK-788’s is about 25%. And why does that matter? Because presumably it will influence not just the response rate but the duration of response, the PFS, et cetera.

DR LOVE: In terms of benefit, is it kind of similar in waterfall plot, duration of response, to what you see with the more conventional ones?

DR CAMIDGE: It’s not quite there yet. We’re running about a 40% to 50% response rate. Median PFS again here was influenced by the presence or absence of brain mets, but if you didn’t have brain mets it was about 8 months. It’s getting there. It’s not quite the same as saying, “I’ve got an L858R, and I’m going to go on osimertinib.” It’s not quite that good yet. I think 788 is looking the more promising of the pack. Poziotinib, I think, definitely got people excited about the field, but lots of challenges in terms of its tolerability, and at least when they’ve moved it outside from MD Anderson and people are using it in the real world, its response rate seems to be dropping, too.

DR LOVE: In terms of tolerability, is it kind of like afatinib? In other words, worse than erlotinib and way worse than osimertinib type thing? Or where does it fit in?

DR CAMIDGE: I would say TAK-788 is probably similar to afatinib. I would say poziotinib is worse than afatinib.

DR LOVE: Hmm. And in terms of the old IPASS bar, do you think this drug or any of these drugs are ready for first-line therapy?

DR CAMIDGE: When you’ve got about a 40% to 50% response rate, that’s kind of on the fence as to whether that’s really better than first-line chemotherapy or chemoimmunotherapy.

DR LOVE: Yes, I’ve heard people talking about 60% as the bar they want to hit.

DR CAMIDGE: Yes, exactly, and I don’t know that we’re really getting it. And one of the things is, as I hinted, some of these exon 20 insertions may be different, and so we think we’re comparing like with like, but even if you look in the preclinical data — for example, there’s something called a D770 insertion versus a V774 insertion. One is nearly twice as sensitive as the other one. And so these things could influence that response rate. It might influence the group that you want to use these first line.

At the moment, I’m tending to use them second line, unless you’ve got a patient who, for whatever reason, you think chemo is not a good deal. But this is not a free lunch either, in terms of its side effects.

Incidence of NTRK alterations and treatment for patients with NTRK alterations

DR LOVE: I wanted to ask you also about NTRK fusions. We just did a survey of 77 general medical oncologists. It was really looking at the kind of assays that they’re doing. This is actually breast, colon and prostate cancer. But I was amazed at the volume of NGS that was going on. And they’re finding NTRK. They’re finding MSI. They’re finding all kinds of PARP stuff. I was really amazed what’s going on in practice.

I asked you if you had anybody who had an NTRK fusion. You said no, but maybe you can talk a little bit about where we are right now with that.

DR CAMIDGE: Two things. I think we predicted, maybe back in 2010, 2011, that once you got past 4 anilides, you got to the tipping point in health economics. That actually doing a panel was cheaper than doing individual anilides. We are totally there with lung cancer. We are getting there with many other tumor types. That doesn’t surprise me that people are checking the box and say, “Go look for everything in your panel.”

I think we will have to start to know what’s in the panel. There’s a snob value. I don’t care if you have 400 genes or 20 genes. I just need to make sure you include the genes which are actually driver oncogenes. Most of that’s just kind of baggage. Like all the things on a new car that you never actually use.

I think what we’ve got here is with NTRK, I think what I told you is that we are 1 of the 2 sites, along with Dana-Farber, that first described NTRK gene rearrangements in non-small cell lung cancer. And so we have been very proactively looking for it, even from the get-go. Initially we developed a FISH assay, and then it became part of our next-generation sequencing panel.

But I have to tell you, in 3 or 4 years we have not found another NTRK gene rearrangement in lung cancer. We found them in GBM. We found them in a few other kind of rare solid tumors, and they’ve gone into the study. But these things in lung cancer really are fantastically rare. I mean, I remember Lecia Sequist saying it’s 1 per career.

DR LOVE: Wow. From your point of view, what is appropriate therapy for these patients, whether it’s lung cancer or otherwise?

DR CAMIDGE: The irony is, we have this thing that you’re going to see 1 in your career, and you’ve got 2 drugs. You’ve got larotrectinib and entrectinib, and they both have 60% to 80% response rates. You’re kind of spoiled for choice. Larotrectinib got licensed slightly earlier. I don’t know there’s much to choose between the 2. I would try one, and if it didn’t work you’ve got the other one as a backup.

Case: A man in his mid-60s, a heavy smoker, with metastatic squamous cell lung carcinoma and MET amplification receives carboplatin/paclitaxel/pembrolizumab as first-line therapy

DR LOVE: Let’s talk about the other patients that you wanted to discuss here today, beginning with your 64-year-old man.

DR CAMIDGE: I guess I brought this one in because I wanted to make a point. Almost like our first patient who was the heavy smoker who still has an ALK rearrangement, 64-year-old guy, 22 pack-year smoker, he presents with a seizure, actually in your neck of the woods, in Florida. He was visiting family there. And he has a 3-cm lesion in his left occipital lobe, a 1-cm lesion in his left frontal lobe and about a 1 centimeter in his left parietal lobe. They resect the occipital lesion, and it is TTF1-negative, P63-negative, P40-negative, so negative for almost everything. But they call it squamous based on the physical appearance of it. And so at least where he was in Florida, they said, “There’s no need — you’re a smoker. You’ve got squamous cancer. We’re not going to do any molecular testing.”

He comes back to Colorado, which is where he lives, and he got a PET scan, and in addition to the stuff in his brain he’s got some hilar activity, 4R lymph node activity and with no CT correlate there’s some activity in his left femur. Actually, interestingly enough, it wasn’t clear where the primary was. It was probably tangled up with the hilar mass.

He gets radiotherapy to the resection cavity and SRS to the 2 of the lesions in his brain. And the molecular testing comes back, and we do 2 next-generation sequencing panels. One is an RNA extraction, and one is a DNA extraction. One is to look for MET exon 14 and gene fusions. One is to look for point mutations. But we also still do MET FISH testing, because MET gene copy number is still the only thing that is a true gene copy number-driven driver. And I truly do not believe next-generation sequencing is the best way to pick that up. And on FISH testing he had a MET/CEP7 ratio of 5, which is high. And that was not picked up by the NGS, but there it was on the FISH testing.

We started on carbo/paclitaxel/pembro. He did pretty good on it. And then because he really had, outside of the stuff in his brain, fairly oligometastatic disease, we did consolidation radiotherapy after about 6 or 7 cycles. And then he’s done fine, but a bit like the other lady, he’s now starting to crop lesions in his brain. And the question is, where do we go next? His disease outside of his brain is well controlled. Maybe there’s an issue in terms of if he was responding to the chemotherapy or the immunotherapy, it’s not getting up there.

I know he has potentially MET as a driver, so a high-level MET amplification ratio has about a 40% to 60% response rate to a MET inhibitor like crizotinib. But crizotinib’s not going to get into his brain. I have solicited for single-patient compassionate access to tepotinib, which he is going to get, and we’ll see how it goes.

DR LOVE: Tepotinib. What’s that?

DR CAMIDGE: There are 2 next-generation MET inhibitors that are really chasing down to be the first one to get a MET exon 14 skip mutation license, and both of them have good activity, and both of them have some evidence of CNS activity too. One is capmatinib, and the other is tepotinib. I always remember that because it’s like a teapot, tepotinib.

DR LOVE: Right.

DR CAMIDGE: And I don’t quite know why I went for tepotinib over capmatinib, but that’s the one that we’re going to try and get single-patient access to.

DR LOVE: Interesting. Also interesting, it looks like he had a very good response.

DR CAMIDGE: Yes, yes, he did.

Choice of chemotherapy for patients with metastatic squamous cell lung cancer and targetable alterations

DR LOVE: And that also brings up the question of taxanes. You have both nab paclitaxel and paclitaxel approved as first-line treatment for squamous combined with pembro and carbo. When we poll general medical oncologists in practice, we know that nab is used much more by investigators — perhaps up to half of patients in general medical oncology practice might get nab.

So there is an approval and data with nab and paclitaxel. How do you make that decision?

DR CAMIDGE: I have to be honest, I don’t use it very much, nab paclitaxel, unless somebody’s having infusion reactions to paclitaxel. I tend to be sticking with the original taxanes.

DR LOVE: And you also brought up the issue of testing in patients with squamous. This patient seems a little unusual. But what is your recommendation in that regard?

DR CAMIDGE: I got an email from a lady in New York who’s ALK-positive, card carrying ALK- positive, proven by a next-generation sequencing, responding to ALK inhibitors. But the histology is quoted as squamous. And so her doctor is saying, “When you get onto chemo, I’m going to give you a taxane. I can’t give you pemetrexed.” And I said, “That’s nonsense. Pemetrexed is the chemotherapy that works best in ALK.” Nobody’s done the study as to pemetrexed in ALK-positive squamous cancer, so I think we have to be careful about the rules that we put in place that are not based on anything. In terms of tolerability, in terms of efficacy, given this exaggerated sensitivity of pemetrexed in all of the gene rearrangement subtypes of lung cancer, I suggest she try that rather than jumping straight into paclitaxel or whatever they were going to give her.

Efficacy of the selective MET inhibitors tepotinib and capmatinib for patients with advanced NSCLC and MET exon 14 skipping mutations

DR LOVE: You have another patient here, but it looks like it’s another MET case, right?

DR CAMIDGE: It is, so the first one was MET amplification.

DR LOVE: Right.

DR CAMIDGE: This was a MET exon 14 skip mutation. And I wanted to bring this up, because I said there’re these 2 drugs, capmatinib and tepotinib, and a few others that are chasing down the MET exon 14 first-line signal. And I think there’s something really interesting going on here. It’s almost kind of like what we talked about with the exon 20 drugs. The response rates that they’re seeing are all in the 40% to 50% range. They’re not in the 60% to 70% range that we’re used to with EGFR and ALK. And they’re sort of ignoring that elephant in the room and rushing ahead and saying, “We can figure out whether this is a second-line drug or a first-line drug.”

But I actually think there’s something much more interesting going on here. There’s a really nice, I think, 2 or 3 papers now, looking at patients with MET exon 14 skip mutations and (a) showing some of them do not express the protein, which is in theory the mechanism of activation. MET exon 14 skip mutations remove something called the Cbl binding domain. You increase the half-life of the protein, so it accumulates on the surface of the cell and it autoaggregates and starts signaling.

But some of these people don’t have protein. Clearly that is not the driver. And some people, maybe about 10%, have a clearly identifiable other driver. Again, MET exon 14 was attempting to be a driver and somehow it got subverted, and the cancer used something else. Mostly we’ve seen KRAS mutations, but we’ve also seen EGFR mutations and ALK.

And that means maybe some of that 40% response rate is your denominator, is stuffed with people who actually aren’t sensitive to MET inhibitors because they don’t make the protein or they’ve got another driver oncogene. And I think when you look at the studies, capmatinib made a big deal out of the fact that their first-line study had a 70% response rate, but in the second line it had a 40% response rate like everyone else. And if you look at tepotinib, its response rate is 40% or 50% whether it’s in the first line or second line. It’s not that capmatinib is somehow a better drug in these small studies, I think they just got lucky and had a few more of the people who are truly MET driven.

And the reason that matters is, if somebody comes along with a much more definitive predictive biomarker, then the whole field gets shaken up. And I think that’s what’s going to come. I think you should get MET exon 14 genetic testing, and then you should at the very least confirm it with MET protein expression. Because you don’t want to waste someone’s time with a treatment that’s never going to work, if that’s not actually their driver.

DR LOVE: Do you think that that’s a feasible goal?

DR CAMIDGE: Yes. I mean, I think what will happen is, the companies who are going to be first to market will ignore me, which is totally fine. Their goal is to get their product licensed and to make money from it. Whoever is Johnny-come-lately with their third to market probably needs to invest a lot of money in developing a more specific predictive biomarker, because then they’ll be able to claim their response rate is 70%. It’s not because their drug’s any better, they’re just giving it to the right people.

Imagine we go back in the time machine and you’re developing a drug for EGFR, and your predictive biomarker is EGFR FISH amplification. And then I’m coming out later and saying mine is an EGFR mutation, so it may be 1 gets licensed, has a 30% response rate, and I come in with a better diagnostic, and I have a 70% response rate.

DR LOVE: I actually remember that happening.

DR CAMIDGE: Yes.

DR LOVE: Right?

DR CAMIDGE: It wasn’t that long ago.

DR LOVE: That was the case not too long ago, exactly what you said, right? I think there was — wasn’t the adjuvant trial, I forgot the name of it, was based on amplification of EGFR?

DR CAMIDGE: That was the RADIANT trial, yes it was.

DR LOVE: Right.

DR CAMIDGE: Good memory.

Prognosis and treatment for patients with NSCLC and KRAS mutations

DR LOVE: I’ve got to ask you about KRAS.

DR CAMIDGE: That’s another great example. That fits in perfectly, because what are we seeing? On the one level it’s almost like we’re seeing with poziotinib, proof of principle. We’re getting responses. Suddenly with KRAS G12C, we’re getting responses. This is very exciting. But then it’s starting to look a bit like more the MET exon 14, we’re getting response rates of 40% to 50%. Some of the people are blowing through it. Some people are not responding at all.

I don’t think KRAS G12C alone is enough of a predictive biomarker. There’re presumably contextual mutations that say whether this is enough to inhibit with the drug. It might be enough to get the drug licensed, but that’s the beginning of the research. It’s not the end of it.

DR LOVE: Yes, as you were chatting about that, I was — in terms of prediction of treatment benefit with these agents, first of all, in terms of KRAS in general, I don’t know, I’ve always thought about these patients just having these terrible tumors that nothing is going to work for. Is it really just another activating mutation, just one that we don’t have a good drug for?

DR CAMIDGE: No, it’s more complex than that. There’s clearly heterogeneity. If you sit in a lung cancer clinic, yes, the general impression is, KRAS patients do badly, but there will be some patients who do incredibly well. And some patients, even amongst the badly, who do super, super badly.

We actually published a paper together with Vanderbilt even before these kind of comutations came out, just putting people into clinical syndromes. And we had a syndrome X, which was just, like, the worst thing. You had a KRAS mutation, and you also had to have a heavy smoking history, and you usually had soft tissue metastases. And their median survival was measured in weeks. The alternative was a group of people who tended to have a KRAS mutation, tended to have a lower smoking status and tended not to have soft tissue metastases. And they could have amazing responses to pemetrexed. They could be alive for years. KRAS alone is not enough to define a population as uniform.

DR LOVE: Interesting.

Predictors of benefit with immunotherapy for patients with NSCLC

DR LOVE: I was also thinking, when you were talking about predictors of treatment benefit to targeted therapy, in terms of the vexing problem of predictors of benefit of immunotherapy, you mention tumor mutation burden, PD-L1. Any personal pet hypothesis about where this might end?

DR CAMIDGE: There’s some nice data that I think I’ve started to see presented in some studies — I’ve seen Naiyer Rizvi present some of this — is take your population who is PD-L1-positive, but then the easiest way to now fragment it is based on the presence and the pattern of infiltration of CD8-positive lymphocytes. Because now we know there’s a group of people who have no TILs, or others are there and they’re clustered at the edge of the tumor, or others where they’re in there but not doing anything. And these undoubtedly are telling us something about the biology in other checkpoints that may be involved. And I think that’s how we’re going to need to present these as the very first thing.

And we wrote a paper with Keith Kerr, who’s a pathologist in Aberdeen in Scotland, and maybe this is his bias, but he was like, “There’s so much more we could do with basic immunohistochemistry to actually define these into different groups, because they look different down the microscope.”

DR LOVE: That’s really interesting.

Ongoing investigation of anti-PD-1/PD-L1 and anti-CTLA-4 antibodies for NSCLC

DR LOVE: Any thoughts about where combined immunotherapy — the regimen we’ve heard the most about has been ipi/nivo, but there are others out there — in terms of where that’s heading? And again, thinking about predictors, what kind of profile might match up with that?

DR CAMIDGE: Ipi/nivo, I think, is going to be a challenge in lung cancer. And you can’t fault them for not doing a good study and not having all of the appropriate controls in there. The problem is that we have a fair bit of evidence, at least in lung cancer, that nivolumab is not as good as pembrolizumab. It has a series of negative first-line studies, and pembrolizumab has a series of positive first-line studies.

The problem there is, it now means that because nivo is part of that backbone, this then becomes an uninterpretable regime. Is nivo/ipi just making nivo alone as good as pembro? I mean, that’s the challenge. I think it opens up the field to saying, “Can we add a CTLA4 inhibitor to our preferred PD-1 — or PD-L1 inhibitor,” which wouldn’t be nivolumab. And that’s interesting. Again, let’s be honest: These additional checkpoints are not going to work in everybody. The immunotherapy field has been reluctant to consider the idea of personalized medicine until literally it was hammering on the door saying, “Look guys, you spent $100 million and you have a negative study,” and then suddenly they’re interested.

We will need predictive biomarkers to say, “Who needs CTLA-4? Who needs LAG-3? Who needs TIM-3?” And that has to come.

DR LOVE: That’s a really interesting point, though, about the nivo. I guess your bringing up the question of nivo makes me wonder, do you think other CTLA-4/PD-1 combinations maybe would be more successful than ipi/nivo? And again, I know there’s some data coming out looking at durvalumab and I think it’s tremelimumab. Any thoughts about that?

DR CAMIDGE: I mean, there is some small Phase II studies going on, but I think what people are starting to look at is, “Okay, add 1 of these things into pembrolizumab.” That’s what we really want to see, is ipi/nivo — if I can be as uncharitable as possible, is ipi/nivo just a way of making inadequate nivo almost as good as pembro? I mean, and cost more and give you more side effects? That’s the least charitable interpretation.

DR LOVE: Are there studies looking at other CTLA-4 plus pembro?

DR CAMIDGE: Yes. There’re small studies at present. They’re not registrational studies, but they’re starting to explore that. But even if they’re positive, it’s not that adding in a CTLA-4 is going to work in everybody. And I think if you said your problem is you don’t have enough TILs, then CTLA-4 makes — fine. If you’ve got a tumor stuffed full of TILs and they’re just not working because of some other checkpoint, CTLA-4 is probably not going to do anything.