Management of NSCLC with ALK or ROS1 rearrangements

DR CAMIDGE: Hi, I’m Ross Camidge. I’m the Director of Thoracic Oncology at the University of Colorado Cancer Center, and I’m delighted to give you this update on the management of advanced ALK and ROS1 rearranged non-small cell lung cancer.

So, let’s start off with a conundrum which is facing all of us in the clinic, so a patient with advanced ALK-positive lung cancer walked through the door. What is our first-line treatment decision these days? Let’s have a real case.

So this is a 67-year-old Caucasian female. She was a biologist, so she understands some of the conversations that we’re having. She’s a nonsmoker. In April 2014, she had an incidentally found Stage IA cancer, which was resected. She goes through the usual surveillance. She eventually gets to annual surveillance, and then because of COVID, her April annual scan gets delayed until August, probably not a big, meaningful delay. But unfortunately, on that August scan, suddenly on the same side where her original cancer was, there is now a medial right pleural-based mass abutting the mediastinum. There is some lower right pleural-based masses and some nodules that you can see in the fissure.

PET-CT scan shows that they’re all avid, but she has no additional sites of disease, so this is an M1A kind of scenario, and a brain scan is clear. She has no symptoms, so this comes as a bolt of the blue. She gets a biopsy of 1 of these pleural-based masses and it shows by next-generation sequencing that it's EML4-ALK-poisitve. So what do you do?

Well, life was easy until really relatively recently. So, 2 of the next generation drugs, alectinib and brigatinib, had both gone head-to-head against crizotinib. And, although they had very different outcomes in the post-crizotinib setting, the median progression-free survival for brigatinib is like 15 or 16 months; alectinib, 7 to 9 months. In the first-line setting, they looked almost identical. I mean it's quite incredible. If you look at different endpoints, whether it's the investigator-assessed or the independent radiology review assessed, or whether it's the hazard ratio or the median or a landmark estimate, they’re essentially identical. There was a subtle difference in the median PFS point estimate by the investigator that was somewhat high with alectinib, but, in general, these things are identical. So those differences post-crizotinib didn’t turn out to be different.

So, you essentially had a head-to-head choice. Both are pretty well tolerated. Subtle differences in terms of the number of pills you take, the side effect profiles, but 90% of people are taking alectinib, maybe 10% were taking brigatinib in this setting.

And then 2 other trials came out, both in 2020. So this was eXalt3. eXalt3 is a different drug, currently doesn’t have a license in the USA, called ensartinib, and it was compared head-to-head against crizotinib. In the post-crizotinib setting it looked almost exactly the same as alectinib in terms of median progression-free survival and there was nothing really outstanding here. Again, it was a positive study. The outcomes again looked very, very similar to those other studies.

So in different parts of the world, maybe this will be priced differently; maybe it will capture part of the market based on that, but this wasn’t shaking things up.

What shook things up was when the CROWN study came out. Now the CROWN study is lorlatinib. Lorlatinib is a third generation ALK inhibitor. It has quite phenomenal CNS penetration. It has broad coverage of many of the ALK resistance mutations that can occur post-crizotinib and post- some of these other next generation inhibitors, and we didn’t know what was going to happen. But here, you can see that the hazard ratio wasn’t running in the kind of .49, .5 range. It was .28. Now that's not just a chance effect. That's a real difference. You can look at the 12-month PFS rate and you can see that, actually in the experimental arm it was maybe about 10% better than you would expect. The control arm was maybe about 10% worse, so the hazard ratio is a relative term. So you can start to debate why is this different? But then you very quickly have to get down to what do you do with this data.

So, why is it different? Well, I don’t know if that additional mutation coverage is the reason. We saw that the additional mutation coverage of brigatinib didn’t really turn into a difference in the first-line setting. Maybe the cancer just finds a way around it. I wonder if maybe it's that, but maybe it's also this CNS effect of lorlatinib, which is so prominent. Although, what’s really interesting is if you look at the subgroups by whether they did or didn’t have brain metastases at baseline, those who did have brain metastases at baseline clearly had a phenomenal hazard ratio, 0.2. But that's actually very similar to what brigatinib had in the ALTA 1L study.

What’s really surprising is the very low hazard ratio, .32, for those who didn’t have baseline brain metastases, and that's about 70% of the population in these studies.

So you raise the idea that maybe in those with an intact blood brain barrier, those who don’t have brain metastases at baseline, that greater penetrance of lorlatinib allows it to really nail the CNS protection effect. And I don’t know. We can argue about why it's different. But the real question is, what do you do about it?

Well, we don’t have a median PFS of the lorlatinib and CROWN yet, but it clearly looks like it's a very active drug. But, even if it has a median progression-free survival of 3 years or even 4 years, you have to remember that the patients with ALK-positive cancer can live for years now. In our own series, the median overall survival was nearly 7 years. And I have multiple patients who are decades-long survivors with Stage IV disease.

So we are in a brave new world of oncology where our concept of PFS must be king because that's really very close to overall survival, is disappearing. We are in the situation where we probably have to think about the overall trajectory of somebody’s life, because it's about getting perfect control of disease and minimal side effects. And the only reason why I’m raising that here and why you wouldn’t immediately reach for lorlatinib is because lorlatinib comes with some baggage.

So the side effects of lorlatinib are quite unique. In this study where you look in the tornado plot versus crizotinib, you can see that something like 70 to 80% of patients will have cholesterol or triglyceride abnormalities that will actually need treatment. So it's not just a sort of laboratory abnormality, you actually have to put these people on statins.

The other things that you could notice are weight increase. Now something like 40% of people get a weight increase, of whom half are getting Grade 3 weight increases; that's about a 20% increase. So if you weigh 150 pounds, that's a 30-pound weight increase. This is a big deal. You can get peripheral neuropathy. But one of the things that I think most people are concerned about with lorlatinib is its higher cognitive function side effects.

Now they break them down in this which makes them look less common then they are, but if you add together memory, cognition effects, mood effects — people can get depressed, people can get anxious, people can get rage — and they also have a speech abnormality. They will slow their speech. If you add all of those things together, probably nearly half the people are getting this.

Now, that doesn’t always result in a dose reduction. But there’s a concept I think amongst many physicians who treat ALK-positive lung disease, that this is not necessarily a drug which is a free ride. And so, maybe you’re going to need to go on lorlatinib at some point perhaps, but do you really need to start on it, or you’re just spending more time on a less well tolerated drug?

I think one of the things that one should think about is the whole trajectory of the disease. So let’s use a slightly different scenario.

So this is a patient who is a gentleman, Caucasian male, handyman, light ex-smoker. Now, this illustrates a very long history of ALK-positive disease. He was diagnosed in September 2010, nearly 10 years ago. He had a malignant right pleural effusion. It was ALK FISH-positive, that was the diagnostic we used then. He goes on cisplatin, docetaxel and bevacizumab in outside facility, has a great response and has a break from therapy. And then later, maybe a year or so later, he progresses. He goes on crizotinib.

Has a great response for just over a year and then he had some progression. He goes on what was a clinical trial of ceritinib at that point in time. Has a great response. Now ceritinib is not a very well-tolerated drug; he tolerated it fine. And then he had, maybe after somewhat less than 2 years, he had some oligoprogression was zapped with radiotherapy. And then he stayed on the drug, maybe for about another 3 or 4 years, so that oligoprogressive local ablative therapy really kept him going well, until about November 2018. And he progressed.

We rebiopsied. And we found a resistance mutation. Now the licensed drug would be lorlatinib. But you can actually look up these resistance mutations, and this was a rare mutation, but it should have been sensitive to brigatinib, at least in my hands brigatinib is a very well tolerated drug, and we put him on brigatinib. He had another complete response. He’s a complete response lasting for another 2 years.

So again, you start to think about, can you play catch-up with these things later? Do you need to jump onto, perhaps, a less well-tolerated drug from the get-go?

This is the range of different mutations that you can get post- some of these different drugs. But perhaps the thing I really want you to take home here, as we start to talk about the ultimate trajectory of these people, is the elephant in the room isn’t the colored boxes here; it's the gray box. The gray box are second drivers where you can’t find an on-target resistance mutation. That's the next area of research that we really need to know. Because not everybody is going to stay addicted to ALK, and not everybody just needs to go ALK followed by ALK followed by ALK.

You can illustrate this by some nice work on again lorlatinib in later lines of therapy; whereas, if you find an ALK resistance mutation your chance of a response to lorlatinib is about 60%. If you don’t, it's only about 20%. Again, the median progression-free survival is something like 11 months if you can find a mutation; it's only about 5 months if you can’t.

So, these things make a difference. And they also make a difference because it's not just everybody is going to go on lorlatinib and just going to feel better or worse about them responding, we now have identified second drivers.

MET seems to be the winner of the best supporting actor Oscar. MET amplification, MET fusions, MET skip mutations can all function as second drivers. You have to make sure that your diagnostic is looking for these things. And they're actionable. You can stay on an ALK inhibitor and add in a MET inhibitor, and sometimes these people will respond.

Perhaps the most intriguing thing about why that gray box of second drivers is going to take a little while to fill in, is that there are functional second drivers that we can’t find through our usual sequencing mechanisms. This was a case from the World Conference in Lung Cancer, young woman, one of my patients, who was positive for developing acquired resistance on alectinib and then brigatinib. We rebiopsied. No ALK resistance mutation. No MET amplification, fusion, or exon 14 skip mutation. But we were able to grow a cell line. And when you grew the cell line, the cells had huge levels of MET signaling, so we added in crizotinib, that workhorse drug which is also a MET inhibitor. She’s had a complete response for 2 years.

So, identifying those second drivers is going to be one of the next waves of discovery.

So, let me sum up ALK in a nutshell here. So, you’re going to start people on a next-generation ALK inhibitor. That's a given. You get to choose between alectinib, brigatinib, ceritinib, and lorlatinib. Those are the licensed ones. I suspect most people are going to say, look, some people on alectinib can have 2 to 3 years of even longer. It’s a pretty well tolerated drug. I can use brigatinib in the next line. It's a pretty well tolerated drug. I’ve got lorlatinib. It looks good when you go upfront because it delays the time to the next treatment decision, but I don’t know if that's going to matter in the overall pathway of someone who’s going to be alive for nearly a decade, maybe I can play catchup later.

So you’re going to start them on that drug. Sometimes they’ll progress and you’re going to treat with oligoprogressive state with local ablative therapy. But then when you can’t do that, you’re going to want to rebiopsy. If you can find an ALK resistance mutation, choose whichever drug that you are most comfortable with, whichever is best tolerated to go on to, it doesn’t have to be the licensed drug. There are many drugs out there.

Make sure that your diagnostic, if you don’t see an ALK resistant mutation, is broad. That you’re looking for all of those next generation resistance mechanisms, which could be any second driver. We’re getting things coming up. RET rearrangements have come up. MET, as we’ve said, has come up. We are getting the things where we said these things were mutually exclusive, that all breaks down in this setting. If you find one and you have a drug you can add it, think about doing that.

But if you don’t find a resistance mechanism, or you find a mechanism you can’t do anything about, you’re really just going to try an ALK inhibitor, but you should expect it to fail. That's very important. Keep a very close eye on your patients. Because if you don’t know the mechanism of resistance, just jumping onto another ALK inhibitor is likely not to work. And if it fails, then probably you’re going to want to move on to chemotherapy, it should probably be pemetrexed-based chemotherapy. And the big debate there is, do you keep the ALK inhibitor going? I tend to, especially if somebody has brain metastases because it's protecting that part of the body, but that is something which is currently unproven.

Alright, let’s move on to ROS1.

So ROS1, I’m going to show you 2 examples, 1 is to illustrate the first-line decision-making, where a lot of the debate is about the activity of these drugs in the brain. And then I’m going to show you a second-line scenario, where, again, it's about personalized medicine.

So, first case, a 37-year-old Hispanic female. She was an accountant. Again a never smoker. Presents with dyspnea in May of 2013, again prolonged life expectancy with these people. You have to think about the big picture.

CT scan shows a left hilar mass, an AP window lymph node, but a PET scan also shows multiple bone metastases and liver metastases. An MRI of the brain is clear initially.

Liver biopsy shows ROS1 positive by FISH. Again, that was the diagnostic that was used then.

She starts on crizotinib. She does great. But then, something like 2 years later, and then 2 to 3 years after that, she develops solitary brain metastases, treated with stereotactic radiosurgery. The rest of her body is well controlled.

We know that one of the liabilities of crizotinib is it doesn’t get into the brain well. In fact, if you look at all the crizotinib studies, there’s a lot of median progression-free survival jumping around, as you can see here in the red box. It can be as low as 9 months median progression-free survival or it can be as high as nearly 20 months. What’s going on?

Well, one of the things we think may be going on is the different proportion of patients with brain metastases in these studies. The original one, the Phase I that I was involved with, we actually never captured that information. It's never been presented in the public domain, what proportion of those patients have brain metastases. But one might imagine it's relatively low, or a lot of those patients had had whole brain radiotherapy, which essentially sterilizes the central nervous system.

But we know that the rate of brain metastases in ROS1 is almost as high in ALK. Initially, there was some small series that suggested that it was lower, but that's the same series that's probably contributed to that initial Phase I study where there are probably people who had a relatively low rate of brain metastases. So if you look in the bigger picture, it's an issue.

The other thing that can happen, apart from progression in the brain, is you can get on-target resistance mechanisms. The most common one, which can happen in somewhere between 20 and 40%, depending on where you get your data, is called G2032R, but there are multiple other resistance mutations. And, again, you need to know the difference.

Just like we showed for ALK, the elephant in the room is the red box here, where there is no resistance mutation. Again second drivers are probably are real thing with ROS1, and just moving from 1 ROS1 inhibitor to the next, sometimes, if you don’t know what the mechanism is, will fail.

You can look up the sensitivity of these different mutations if you find them to some of the drugs which are either licensed or able to be used off-label as a ROS1 inhibitor. Now, this is a somewhat different way of presenting it compared to some of the traditional ones. Often, people will rank these things just by the IC50, in this particular study by Dimou et al, what they did was they also stratified by the exposure, they looked and adjusted by the average concentration of the drug. Because it’s not just the IC50, if you can drown the cancer in the drug, the IC50 doesn’t matter. So you actually have to factor in the exposure in the individual. So, I actually like this way of presenting it.

What you can see here, if you look across some of these rows, is that some of the common resistance mutations are indeed sensitive to some of the other common drugs. So, for example. D2033N is clearly sensitive to lorlatinib. G2032R importantly is not sensitive to lorlatinib. And it's not sensitive to ceritinib. And it's not sensitive to entrectinib. It needs a different approach. So not all ROS1 resistance mutations are the same and not every patient is resistant because of a ROS1 resistance mutation.

Let’s look at the entrectinib data. So, entrectinib is the other first-line licensed drug for ROS1 next to crizotinib. If I summarize it, the key thing about entrectinib is it's a CNS penetrant crizotinib. It doesn’t really have any additional anti-mutation coverage, but it does get into the brain. You can see here it has a response rate of 55% in a relatively small series. And so, when you start to see its median progression-free survival, so the median’s about 19 months, that looks a lot more like what we saw in the profile 1001 study with crizotinib. So it looks relatively similar. But we know that something like 30% of the patients in this study had brain metastases.

So, again, maybe what’s it's doing is lifting those up to look a lot more like crizotinib in a non-brain met population.

However, the other I want you to look at in this data is the median progression-free survival is still lower in those with brain metastases than those without. So entrectinib is better at getting into the brain, but it's not perfect. So it's still a relative liability.

However, many of the pathways have said, look, you’ve got a choice between crizotinib and entrectinib. Entrectinib gets into the brain. Why don’t we just use that as our go-to drug.

What about side effects? They are slightly different. Entrectinib, as you know, is also licensed as an NTRK inhibitor. That gives you some higher function abnormalities. You can get some taste disturbance and dizziness. But for most patients they’re not deal-breakers.

Let’s move to the second line. So you started them on entrectinib, or you started them on crizotinib. They then progress. So, here’s our same patient. They’ve had multiple new brain lesions turning up. Probably you’re starting to become a little uncomfortable with what’s going on. They were also progressing in the body. Well, what they did was they put them on lorlatinib. It's not a licensed ROS1 inhibitor, but it has ROS1 activity. It's in sitting in the pharmacy for ALK. And what happened was the brain responded beautifully. And that’s probably because the drugs weren’t getting in the brain and just lorlatinib gets in. But the lungs didn’t, the pleural disease progressed.

When we biopsied it, there was no detectable ROS1 mutation. This is a big clue that it's progressing probably because it has a second driver. Now they felt that whole ROS1, there’s ROS1, there’s ROS1, they sent the patient in here for a repotrectinib study. That's one of these next generation drugs that does work in G2032R, and there was no benefit.

So, again, that comes home to the idea that if you don’t know the mechanism of resistance, just simply plying onto the next and the next and the next ROS1 inhibitor can be doomed to failure.

Let me show you some real data. So here’s lorlatinib, lorlatinib in the post-crizotinib setting in ROS1. Response rate is only 26%. Now why? Well, some of these people don’t have a resistance mutation and some of these people have G2032R, which is not sensitive to the lorlatinib. However, lorlatinib is really great at getting into the brain. And the response rate is double that, it's 50% in the brain, and that's probably because this is essentially a treatment-naïve population in the brain.

Let’s look at repotrectinib. This is only available in clinical studies. Here’s the data in those with G2032R, relatively small numbers, but they have a 43% response rate. That looks encouraging.

I want to show you one little thing about repotrectinib. They tried to develop repotrectinib as a first-line drug. I think that's probably a mistake. I think they should probably focus on it in G2032R initially. And the reason I think it might be a mistake is it's a bit like lorlatinib in the ALK-positive population. It carries with it a fair amount of baggage in terms of side effects. But I want to show you one little thing that I want the whole field to become more aware of in terms of how we present toxicity.

So, here’s the way they present the toxicity on repotrectinib, 93 patients treated across multiple different doses. And they show that, yes, you have some taste disturbance. You have some paresthesia, ataxia, muscle weakness. Sometimes patients refer to themselves as having spaghetti-legs. But what I want you to see here is those grades are covered across all 93 patients, but if you look down in the bottom right, something like only 70% of people were treated with that dose and 30% were treated at lower levels.

So when you present all of your toxicity, at all doses combined in a Phase I study, the lower dose levels are diluting your apparent toxicity signal. This is a little bit of sleight of hand. And I think we started to see this more and more.

I would really encourage you to look at this paper that came out in April 2020. And what we did was we looked at the way toxicity was presented in early phase studies in lung cancer. And it's astonishing how poorly standardized it is.

We ended up coming up with a range of suggestions that, if there is a dose escalation element, you should present at the recommended Phase II dose, and not just a combined dose level. That was the issue with repotrectinib.

We also reported that you should really break down the granularity, and we’re hoping is that this will become a sort of minimal disclosure because we’re making decisions on earlier and earlier data and it becomes incredibly important for the optimal management of your patients.

So, thanks very much for listening.

Therapeutic approach to metastatic NSCLC with ALK rearrangements in the first-line setting

DR LOVE: I just want to ask you a couple of follow-up questions about your talk, your 67-year-old lady that you presented in the beginning, what did you end up treating her with? And how long ago was this? Because I’m kind of curious what your typical first-line therapy is.

DR CAMIDGE: She was like 2 weeks ago.

DR LOVE: Okay.

DR CAMIDGE: She was the brand new person. So I gave her alectinib. I mean when the eXalt data came out, you have to think long and hard, and I think the analogy I really want to try and think about is, if you’ve got something that really controls your brain and you delay the time to the next treatment decision, that's good, but you’re paying a price in terms of side effects. And I actually think lorlatinib is almost like the whole brain radiotherapy for ALK inhibitors. And it's really interesting.

DR LOVE: Wow.

DR CAMIDGE: There’s a really cool study that was led by Paul Brown, published in JAMA in 2016, and they did SRS plus or minus whole brain radiotherapy. And they looked at cognitive impairment which was clearly worse. But the hazard ratio, if you used whole brain radiotherapy for intracranial progression, was 0.27. Like remarkably similar to the hazard ratios we’re seeing with lorlatinib. And yet, we don’t use whole brain radiotherapy if we can avoid it. So we know in that setting that we can delay the time to the next treatment decision, but we recognize the cost that we’re paying and lorlatinib makes me feel a little bit like that.

DR LOVE: No, I think it's a really great point. But in general, this sounds — I think if I remember the case — not that unusual of a situation.

Is your typical first-line therapy alectinib, or do you sometimes use anything else, brigatinib? Lorlatinib?

DR CAMIDGE: Obviously, I was involved in the development of both, but I would actually use alectinib because you can go on alectinib and maybe you can have 5 years. And who knows what we're going to know in 5-years’ time. And if it's well tolerated, why not stick with it? If you progress, I’m going to try and find out why. If I can find a mutation — a better tolerated drug than lorlatinib can cover, I’ll do that.

So my case was a guy who had a mutation that could be covered by brigatinib. He has no side effects on brigatinib. So I think it's about personalized medicine.

DR LOVE: Well, I’m just trying to understand the alectinib/brigatinib decision. What was the reason — why do you choose alectinib over brigatinib as first line?

DR CAMIDGE: Because I think I’ve got clear evidence of activity with brigatinib, post-alectinib, whereas, I don’t really have it the other way around.

DR LOVE: Wow, interesting.

DR LOVE: So, just to be clear though, because I think the last time I talked to you, you were saying you were using brigatinib first line.

DR CAMIDGE: No. I mean, I think —

DR LOVE: Brigatinib first line? You never did?

DR CAMIDGE: Only in the clinical trials. I mean I think I really, if brigatinib had good tolerability and amazing efficacy, maybe I would. But life became easier. It became, alectinib and brigatinib were very simple. I didn’t have to choose between my children as such.

And I think one of the really important things is — so, Tom Stinchcombe and I just published a paper, showing that the activity of brigatinib post-next generation inhibitors, has a 40% response rate. Lorlatinib in the same setting has a 45% response rate.

But the key thing, and I actually got this from Zosia Pietroska, is that's not telling you that it's 40% benefit in everybody. What it's telling you is there’s a proportion of people who don’t benefit and a proportion of people who do. And you can actually out figure out who they are by rebiopsying them.

Case: A man in his mid-60s presents with NSCLC with an ALK rearrangement and a solitary brain metastasis

DR LOVE: Actually, I want to start out with a case that I heard from a general medical oncologist. And he presented this case. He said, “I don’t know what to do.” And I said, “Hey, guess who I’m interviewing, I’m going to ask him” because it is an ALK case. And here’s the case, it really is a real-time case, but I think it's interesting from a teaching point of view.

So a 66-year-old man, presents with metastatic non-small cell, pleural effusion. Mediastinal adenopathy and a solitary brain met. On bronchoscopy, he’s got adenocarcinoma. Not enough tissue for NGS. And he gets started on platinum/pemetrexed, RT to the brain lesion. Meanwhile, he gets a biopsy and is found to have ALK. And so, I’m just going to stop at this point, before we go on, because I think there might be an interesting teaching point here. Which is, he didn’t give the patient a checkpoint inhibitor, he had a high PD-L1. But he held off on the checkpoint inhibitor until he got tissue. And it turned out it kind of was a good idea because he had a targetable mutation, with this idea you don’t want to have somebody on a checkpoint inhibitor and then find out you want to give them targeted therapy.

First of all, any comment about that?

DR CAMIDGE: So I think that's a really progressive approach. The other thing that would have made it perfect where he could also have held off on the stereotactic radiotherapy to the brain lesion because it as solitary, presumably they didn’t have any side effects. You could have waited to see if they had a targetable abnormality with a CNS penetrant drug, too.

DR LOVE: Yeah, that's a really good point because I guess if you knew the patient had ALK — I mean do you usually do stereotactic or you usually just give them targeted therapy?

DR CAMIDGE: If I knew they had ALK — I mean we have an embarrassment of riches in terms of CNS penetrant drugs. So I would give him the drug. They will be responding in the brain before they’ve gone into the Radiation Oncology department.

DR LOVE: Right.

So, interestingly, the patient gets put on alectinib, and then progresses in 13 months, so it sounds pretty short to me. Actually, I think he had hepatic mets at this point. And you can see what was seen on the biopsy that he had, in addition to ALK, this EGFR D761N, TP53, etc. High PD-L1. And for better or worse, he was switched to lorlatinib and had a lot of problems with it. A lot of psychologic/psychiatric personality change. Major stuff, behavioral stuff. Wife, very, very upset. They stopped lorlatinib, the patient goes back completely to normal. But the patient is progressing. And they get a liquid biopsy.

And this is the main thing, that he had this question about.

So now he’s progressing after getting alectinib and not tolerating lorlatinib. So he gets a liquid biopsy —

DR CAMIDGE: Neil, do you know — Neil do you know if he was responding to the lorlatinib?

DR LOVE: I believe so.

DR LOVE: So, he wants to know what to do. What are your thoughts about what’s going on?

DR CAMIDGE: So what we’re looking at is the variant allele fraction in circulating free DNA. And what I don’t have is the date for when the baseline was done in the second one. But what you can see is the ALK fusion has gone down a little bit.

DR LOVE: Oh, so this is a before and after thing?

DR CAMIDGE: That's what I’m assuming. So we don’t have the dates written down there. So the question is when did he do — so he must have done a baseline. Because we’ve got a variant level —

DR LOVE: This is previous biopsy –

DR CAMIDGE: — and then we’ve got not detected.

DR LOVE: Right. So he sent it to the same place and they reported both. Okay, great.

DR CAMIDGE: Because he has that EGFR mutation that you described previously that he has a prior biopsy showing the EGFR mutation was actually cfDNA, and not a tissue biopsy. And I think what I would take away from this is probably that the patient was responding to the lorlatinib; they were on it for 9 months. And the fact that they didn’t find a mutation, that falls into that category where you’re worried it won’t work, but you’re not guaranteed it won’t work. In Alice’s study, there was a 20% response rate and the median PFS was about 5 months. Because sometimes you can miss it. And I think picking fusions up in the blood and picking mutations up in the blood, again we all know if you don’t see it, it doesn’t mean it's not there.

I think the other stuff, those other mutations, they're just noise. They mean absolutely nothing. So they’re not mechanisms of resistance.

DR LOVE: So what specific kind of resistance mutations are you looking for where there are drugs that you can use?

DR CAMIDGE: So you would be looking for an ALK mutation, which he doesn’t appear to have detected. And yet, if he did respond to the lorlatinib, one assumes that he may have had one, just that it wasn’t showing up in the blood.

And then one of the issues with cfDNA is, you’re just sampling every piece of DNA in the blood. There’s an algorithm to try and filter out the stuff from normal cells or from subclones of hematopoetic cells. But they can fall through. What you’re looking for is a mutation that is a driver abnormality.

So, a mutation that occurs not in the kinase domain or one that you’ve never heard of, then these don’t mean anything. You can see where it says synonymous alteration. So that didn’t change the significance. You can see where it says variant of unknown significance, that's a polymorphism.

So this is just noise. So, essentially, we don’t know the mechanism of resistance in this person. And what you should do is you should actually biopsy the tissue.

DR LOVE: So you think that would be worth doing, if it's possible?

DR CAMIDGE: Yes, because we’ve got somebody who probably had an on-target resistance mutation, that would have happened in about 40 to 50% of the cases post-alectinib. The fact that they were on the lorlatinib for 9 months would suggest that maybe they derived some benefit. The fact that the variant allele frequency of that ALK fusion went from 0.2% to not detected, suggests maybe there was some benefit. But, if things are now growing, the question is why? And it's clearly not showing up in the blood. Maybe we’re keeping a very close eye on this patient, so it's a relatively small volume of progressing disease. But what we want to be looking for here is either a compound mutation in ALK that lorlatinib will hit, rare ALK mutations that lorlatinib won’t hit, like L1198F, or a second driver.

DR LOVE: Such as?

DR CAMIDGE: MET fusions, MET amplification, RET rearrangements. Essentially, all those things that we were taught a few years ago that are mutually exclusive, are not when you totally perturb the system.

DR LOVE: And just at a basic point of view, if you look at a report like that, and certainly understand the limitations of liquid and whether it's there or not, but is there something that you see that, for example, that would tell you that these 2 MET alterations that I see there, I don’t know what they mean, how you differentiate that from somebody who has a MET that you can target?

DR CAMIDGE: So, I mean the great thing is that we don’t have to memorize these things. You just take that mutation; you stick into Google and you see what comes up. Many of them will be polymorphisms. You’re trying to figure out if this somehow is being described as able to drive a cancer cell all by itself, in which case it can have a role as a second driver.

If it's something that you’ve never heard of, and no one else has ever heard of, please don't start adding in drugs just because on the basis of a cfDNA test. It's going to go wrong.

DR LOVE: What about if there isn’t a specific driver, if the patient is clinically stable, would it be worth considering some kind of targeted therapy, or chemotherapy? And if you went for chemotherapy — you mentioned before — in terms of preference for drugs and how you handle PD-1 or checkpoint inhibitors?

DR CAMIDGE: So in this person, let’s say we didn’t find any — let’s say we rebiopsied whatever was progressing, and we found nothing. Maybe we’ll find the EML4- ALK, but it doesn’t appear to be altered and we don’t find a second driver. And let’s say it's not suitable for local ablative therapy with focused radiation. Then we’ve really got two choices. The most prominent one is you’re going to go onto chemotherapy. This patient started on platinum/pemetrexed. We don’t know how long they were on it for, probably not very long. So we don’t know if they’re in the group that's super-sensitive. But that's where I’d start. And then you’d have a debate as to, do you keep the ALK inhibitor going?

I probably would because this is a patient who’s had brain metastases. But I probably wouldn’t make it lorlatinib, because they don’t seem to be tolerating it well. So, maybe I’d stick them back on the alectinib and add in platinum/pemetrexed.

The other thing you can do, and this is in the Hail Mary pass category, is I told you about a case where there was a patient who — we could only figure out that MET was their second driver because we could grow the cells. So we know there are second driver pathways that exist that we cannot detect by our current sequencing methodology. Because it's not turned on by a mutation or an amplification or a gene fusion. And the problem is we have no idea what the frequency with which this can occur.

Now, in my case, because we’ve grown the cell lines, we could add in a MET inhibitor and she’s done great for 2 years. And I do have one case from a patient in Tel Aviv who — they were desperate, and I said look, if you’ve got nothing else to do just add in a MET inhibitor. And of course, they responded. So now my N is 2.

DR LOVE: Wow.

DR CAMIDGE: And very happy son. It feels like how you used to use erlotinib when the patient didn’t have an EGFR mutation. But I don’t want to overemphasize that. You could do that 99 times out of 100 and nothing would happen. It just proves again that there are ways the pathways are turned on that we do not yet have a means of assessing.

DR LOVE: One more question I want to ask you about. Again, getting back to your 67-year-old who presented a couple of weeks ago for first-line therapy. Just kind of curious. I note in her history that she had a Stage IA cancer removed 6 years ago. I would guess she didn’t have metabolic profiling then, but I assume if she did, she would have —

DR CAMIDGE: I mean we’ve actually had a case like this before that we’ve discussed, with somebody with a really, really indolent history that 7 or 8 years go by and it's the same molecular driver. Now in this particular patient we couldn’t find the old sample to test.

DR LOVE: So you saw her when she presented with the localized disease?

DR CAMIDGE: No, she’d been seen in the outside institution.

DR LOVE: That's what I figured. I’m talking about when she first presented in 2014 with Stage IA and she had the thing taken out. So my question is, and I’ve been hearing this a lot from medical oncologists is, if a patient like that presented today, if this woman were to present today with a localized IA lesions that had an EML4-ALK, or if she had a Stage II or III, if you could, would you give her adjuvant ALK?

DR CAMIDGE: I wouldn’t. And the two questions which I think are unanswered in the adjuvant targeted therapy are, one, there’s a group of people who are cured who don’t need it. So, can we figure those people out? And the much bigger issue is, the people who aren’t cured, the people who do need it, I don’t know they need to stop it. And that's the thing I worry the most about. I worry that the ADAURA data is immature, and what we’ll really find is — one is sure, we’re overtreating 40% of the people, but the bigger sin might be and then we stop it after 3 years and they’re not cured, and their cancer will come back.

DR LOVE: Interesting. I was doing a symposium.

Roy Herbst was in it, and we were talking about ADAURA, of course, and the issue of having only PFS, and he brought up a really fascinating point which was, and I remember because we were doing a lot of CME, in 2005 the adjuvant trastuzumab trials. Remember that? Orlando, Florida is where — that was where ASCO was in 2005. And everybody went crazy. Change of care immediately. Everybody changed what they did. Even were treating people who were less at risk than were in the trials. And I couldn’t believe, when I go back, it was 50% reduction in recurrence and no effect on survival.

So, oncology, FDA, whatever, I guess we’ve changed our culture. Because we went crazy over that and there’s lots of questions about ADAURA with a much better hazard rate. Any thoughts about that? I’m just pointing out a historical fact, so to speak.

DR CAMIDGE: So that's an antibody as opposed to TKI. I think the important thing is to look at the data that came out in the same ASCO with the Chinese study that had a positive disease-free survival and had zero overall survival benefit. We already know that in the CTONG study, it was just called “adjuvant” — that there was no overall survival benefit. We know in RADIANT –-

DR LOVE: Of what? You’re talking about EGFR?

DR CAMIDGE: Yeah, EGFR mutant, 2 years of gefitinib.

DR LOVE: Hmm.

DR CAMIDGE: It was presented at the same ASCO. It was not brought up by Dave in the discussant. And it was kind of like, guys, we already know that this has no overall survival advantage. Why aren’t you talking about it?

DR LOVE: Really interesting.