-mutant population. By the way, you mentioned something that I want to know your opinion on, because we haven’t talked about what happens once you become resistant to osimertinib. I will leave that for the end. No one wants chemotherapy; it seems that chemotherapy has a bad reputation. Once you start a patient on targeted therapy, everyone wants to remain on targeted therapy. It’s easy: It’s an oral pill, and there are fewer visits to the clinic, specifically at this time with the pandemic. It is the same thing with immunotherapy. You start with immunotherapy, and the patient is doing well, so no one wants to get the chemotherapy. How do we improve on what we have right now? From now on with first-generation and ahead, what do you think as we’re moving forward? 

There are a lot of exciting possibilities. The first and most obvious is that if VEGF inhibitors like bevacizumab and ramucirumab improve outcomes with erlotinib, why not move the patient to a better drug, osimertinib? ECOG is doing a study by Helena Yu of osimertinib plus bevacizumab; that’s a phase 3 trial. Based on what we’ve seen with erlotinib and bevacizumab, I would expect that this trial has a good likelihood of success. We [at The University of Texas MD Anderson Cancer Center] are doing a randomized phase 2 study, and a Japanese group is doing a randomized phase 2 study as well of osimertinib plus ramucirumab. They’re taking the RELAY regimen and swapping osimertinib in for erlotinib. That’s a regimen that we’re excited about as well. There’s a study, the FLAURA2 study. Based on what we saw from the Tata [Memorial Centre in Mumbai, India,] study that you discussed before with the gefitinib combinations, there’s good reason to believe that this would be a positive study.

I’ll mention 1 other approach I’m enthusiastic about. We have a study called NORTHSTAR, and the idea behind the study is this: When patients are initially treated with the tyrosine kinase inhibitor [TKI], even when they have widespread metastatic disease, almost everybody has a good response, and they’re down to just a bit of residual disease. We know that if you didn’t kill it after 3 months, the TKI is never going to kill it, right? Why not use something different to kill it at that point? At the end of 3 months, if they haven’t progressed, we then radiate or resect all the residual disease. It’s a brute force approach, but just because it’s brute force doesn’t mean it’s wrong sometimes. We’re doing the randomization to either continue osimertinib or give osimertinib with this consolidative therapy while continuing the osimertinib. There are 4 different possibilities: adding VEGF, adding chemotherapy, and adding consolidation.

There are other possibilities that are emerging. People are looking at other drugs that they might be able to add with it as well. The early results of trying to add immunotherapy, which is another obvious choice, haven’t been very successful. We know PD-1 inhibitors don’t work great for patients with EGFR mutations, and there were some unexpected toxicities when osimertinib was combined with immunotherapy. That’s 1 that doesn’t look so promising to start with, but we have a number of good options that are in process right now, and I would be surprised if 1 or more of those don’t turn out to be positive. I expect that these are going to be positive studies. This may mean that, a few years from now, we would start with osimertinib, and the patient would get either chemotherapy, a VEGF inhibitor, or a consolidation or some combination of those as well.

We’re going to keep pushing back the PFS [progression-free survival] in that first-line setting. A VEGF inhibitor may add 4 to 6 months, chemotherapy may add 4 to 6 months, and consolidation may add something. I’d love to see us get PFS north of 30 months or longer. That’s where we can start having a big impact, even before we hit resistant disease.

In your opinion, it is chemotherapy, immunotherapy, plus bevacizumab, that is the therapy right now for osimertinib-resistant disease.

Yeah, that’s right. That’s my preference; that is what I recommend. I’ll have to acknowledge this here: We have limited data. We don’t have comparative data.

With the data we have and what we know, we know that VEGF inhibitors are active in 

-mutant tumors, so I prefer the IMpower150 regimen because that was the first-line study that included patients with 

 mutations. If you look at the KEYNOTE-189 study, it did not include EGFR or ALK. If you look at the CheckMate 227 study, it didn’t include those as well. But the IMpower150 study did include 

 mutations. The benefit for the 4-drug regimen with the IMpower150 in the EGFR-mutant subgroup is not huge, but it looked like it had a benefit similar to the other groups. In my hands, acknowledging the limitations of the data set, I recommend the IMpower150 regimen, but this is clearly an area where we need improvement. The questions we raised before are important: Maybe you should keep the EGFR inhibitor going during that time, and maybe we need a different EGFR inhibitor. There are a lot of different things that we could consider there.

To back up what you said, in the initial presentation by Mark Socinski on the IMpower150 study, which was positive for overall survival, I’m going to start with that. In a subsequent update this year, that particular group, the 

-mutant population, continued having the same impact on the overall survival in the control arm.

Yeah, that’s right. My personal approach is to treat with the TKIs for as long as you can, and I tend to be pretty aggressive. If the patient has oligoprogression, and if they develop bone metastases, I radiate it. If they develop a solitary metastasis, I radiate it, resect it, or do whatever I need to. When they are clearly progressing, and you’ve lost control of it with the EGFR inhibitors, if the patient is suitable for the more aggressive regimen, IMpower150 is what I’ve been favoring so far.

There are many things, John, on the EGFR pathway. When you mentioned that you treated your first patient in 1999, that was also when I treated my first patient. I remember that, in May 2019, the patient was still alive. Having a stage IV patient live for 11, close to 12 years already is amazing. We’re looking forward. The only way is through research, and we’re advancing the field specifically. This particular topic has been amazing for all our patients everywhere.

Videos

EGFR-Mutated NSCLC: Looking Ahead 

I’m going a bit deeper on the issue of the RELAY trial because what you said is correct. When you have exon 21, I don’t know what you do in your practice, but it is our duty as physicians to disclose the information to the patient. For example, if you have an exon 21 but no brain metastases, you can easily go to osimertinib. It is different if it’s an exon 21, L858R in which the PFS [progression-free survival], if you compare it with the FLAURA analysis, is significantly lower than what you see in the RELAY trial, which was basically the same number that they did in the analysis population: 19.4 months. We have never seen that for that specific population. It’s something that, in a practice, I would discuss with the patient, and I would obviously go over the toxicity and ask the patient what they would like to do. What do you think about the toxicity described in the RELAY trial, the most cumbersome of which to me seems to be the hypertension, grade 3 level? Tell me your opinion in that regard.

It’s fair to say that it was predictable, it was manageable, and it was consistent. We’ve had a lot of experience with ramucirumab from using it in combination with docetaxel where it’s been approved for a number of years, and what’s been seen there has been consistent. As you noted, the main adverse event noted was hypertension. Other adverse events were pretty mild and pretty easy to manage. There was a little more proteinuria of course, but there was little else that was significantly increased.

One of the concerns is always hemoptysis, which VEGF inhibitors have historically had. A lot of this is from the early work with bevacizumab, where a number of patients had hemoptysis. There was no increase in fatal hemoptysis or life-threatening hemoptysis with ramucirumab. Patients now get a bit more bleeding; it tends to be nose bleeds in my experience. The mucus membranes get a little dried out, so they sometimes get dysphonia, where their throat feels a little dry and their voice box gets dried out a bit. They sometimes get epistaxis, which is persistent nose bleeds. In terms of life-threatening or serious hemoptysis, there was not an increase observed here. The management of hypertension is something that can be done readily; it’s a manageable toxicity.

They had a quality-of-life analysis in the RELAY study, and the addition of ramucirumab did not hurt the quality of life at all. In fact, it maintained it for a longer period of time, so there was good evidence that it didn’t cause a detriment in the quality of life there. It is absolutely manageable.

The ability to get osimertinib after then, to me, is a pretty strong plus. I think very highly of osimertinib. It’s a good drug. We recently reported that, even after patients progress on osimertinib, if they had T790M, you can often radiate 1 or 2 sites and keep them going on the drug, and they can often go for another year. We noticed the PFS2 [second progression-free survival] was a year in that study. We can often keep people going for a long time on osimertinib before they become resistant, and from my perspective, if we have something we can use even before osimertinib, and if you can get a year and a half before you even move to it, that’s an attractive option. It’s valuable to have a couple of different options.

Yes, John. Another thing we learned from that RELAY study was that the incidence of T790M mutation was not affected whatsoever by the addition of ramucirumab. When they analyzed different populations in the trial, the incidence of those patients getting erlotinib alone vs erlotinib and ramucirumab in terms of the incidence of T790M mutation at progression of disease was similar. That’s an important point.

Yeah, that’s right. You do retain the ability to get osimertinib later. If you look at the time until you get chemotherapy, it was delayed in the study, especially for patients who ended up moving to osimertinib afterward. This is why, as you said, Eddie, I have a discussion with patients and understand their priorities. It is sometimes their priority to have the simplest therapy right now and not visit the doctor. In Texas, we often have long drives for people. They may live 6 hours away. If they say, “It’s not feasible for me to get there very often,” then you might say osimertinib. If somebody wants to be aggressive and hold on to their other options, and they can get the intravenous therapy with ramucirumab, it’s absolutely a tenable option to offer it to patients, and it obviously gives us information for building new regimens later.

RELAY Trial Regimen Adverse-Event Overview

What is the rationale behind combining TKI [tyrosine kinase inhibitors] in patients with 

 mutations with VEGF inhibitors, for example. What is behind the dual inhibition, if you can tell us?

It’s a great question. I’ll start by saying that, for a long time with angiogenesis inhibitors, it’s been observed that patients with 

 mutations seem to be particularly sensitive to VEGF inhibitors. It’s not intuitively obvious why that should be the point. 

-mutant patients are not sensitive to VEGF inhibitors. For squamous cell carcinoma, you can go through the list: None of them seems to be particularly sensitive. The most sensitive group we’ve identified has been 

 mutants. In my laboratory [at The University of Texas MD Anderson Cancer Center], we looked at this, and back in 2009 we reported that the combination of EGFR inhibitors and VEGF inhibitors seemed to be particularly active, even in resistant tumors.

As we tried to dissect the mechanism, it looks as though the 

 expression up very high in these tumors. They do it through a pathway called HIF. This is the same pathway in kidney cancer that’s up-regulated for a different reason. In that case, it’s because of 

 mutations. In a similar vein, kidney cancer is sensitive to VEGF inhibition as well, and it’s 1 of the few diseases in which bevacizumab by itself can be active. In that sense, we think there’s an overlapping biology between kidney cancer and 

That’s prompted a number of different studies. Early on, there was a study called BeTa that Roy Herbst led, and the study overall was negative. This was erlotinib plus bevacizumab vs erlotinib. Bevacizumab overall didn’t add benefit, but when they pulled out the 

-mutant subgroup, there was a huge benefit for adding the VEGF inhibitor there. That then led to a couple of randomized studies, and almost all the randomized studies have been consistent in terms of demonstrating at least a PFS [progression-free survival] benefit when you add a VEGF inhibitor to an EGFR inhibitor. There have been a couple of studies now of erlotinib and bevacizumab, 1 led by Takashi Seto and colleagues in Japan, and 1 by a cooperative group in the United States with Thomas Stinchcombe and others. They show a strong trend or a significant improvement in PFS.

For bevacizumab, the improvements haven’t translated into overall survival benefits, and the reasons for that are debated. The studies were not powered for overall survival, so that’s 1 thing that’s important to say. These were positive studies because they were powered for PFS, and PFS was positive. All this supported the notion that 

-mutant tumors seemed to be particularly sensitive to VEGF inhibitors, and these studies had impressive hazard ratios: hazard ratios in the 0.6 range. When you consider chemotherapy plus bevacizumab in the nonsquamous population, the hazard ratio there is closer to the high 0.7s, so this is a much more relative benefit.

That’s now prompted a number of different studies. One of them is called the RELAY study. I can talk about this study. It was erlotinib with ramucirumab vs erlotinib, and this was the first properly powered, randomized phase 3 trial that has readout. Just to remind everyone, ramucirumab is approved for lung cancer. It’s a VEGF receptor 2 antibody. Unlike bevacizumab, which blocks the VEGF ligand, this blocks the receptor. This study was a strongly positive study in terms of its primary end point, which was PFS. The PFS was 19.4 months, and that was about a 7-month improvement compared with erlotinib alone. When you think about that, 19.4 months is about the same as osimertinib. Osimertinib was 18.9 months, so it was similar benefit for an EGFR inhibitor plus a VEGF inhibitor as osimertinib did.

There are a couple of points that are worth noting. It didn’t include patients with brain metastases. If somebody had brain metastases at the start, I would stick with osimertinib because this combination doesn’t have the same track record in brain metastases, but it did show similar benefit if you had L858R or exon 19 deletion. If you look at that L858R population, that’s tougher to treat. They didn’t seem to get as much benefit with osimertinib in the FLAURA study, but they did very well in the RELAY study. They had a very similar benefit here.

The FDA has recently approved this regimen, the erlotinib-ramucirumab regimen. It’s a good thing to give options to patients, and you can think of pluses and minuses for both. On 1 hand, you could say, “Osimertinib is a little better tolerated than erlotinib and ramucirumab, and it’s easier not to get an IV [intravenous], so just get a pill and go home. It’s simpler.” That’s a good argument. The better CNS [central nervous system] activity is a good argument. On the other hand, especially for people who want to be aggressive, they can get the same PFS as osimertinib with the RELAY regimen of erlotinib and ramucirumab. If you get a T790M mutation, which about half the patients do, you can still get osimertinib after that and get both drugs, whereas if you start with osimertinib, you have no other targeted agents after that. That’s an interesting argument.

If you’re looking at what delays the time until you get chemotherapy, you may say that the goal is to avoid chemotherapy, because that’s where patients start to have a significant impact on their quality of life. For a significant number of patients, they can get a similar outcome in the first-line regimen, with the RELAY regimen, and then still get osimertinib for meeting the 10 to 12 months afterward. It is sometimes longer than, then they really push back the time when you need chemotherapy much longer. It’s good to offer options to patients. There are some patients, patients with L858R, patients without brain metastases, who want to be aggressive, and I do discuss this option with them. I’ve just started giving it recently.

Good, I agree. Sometimes, when I listen to you speak, I say, “Maybe John and I went to the same school,” because it seems as if we have the same professor and the same teacher almost aligning on the same concept. For me, in 2020, and perhaps for you and all of us who treat lung cancer, the North East Japan 026 trial presented at ASCO [American Society of Clinical Oncology Annual Meeting]—with an overall survival that you briefly mentioned that was negative—was a disappointment. I saw that study with the impressive PFS that was presented a few years ago. I had hoped it would be positive for overall survival, but that was not the case. When I see that, I say, “Wow.”

Where are we going to put RELAY, which also came this year? There is a difference between bevacizumab and ramucirumab: the mechanism of action is totally different. By blocking the receptor, you block all the ligand, and maybe that is the difference. The other is the dosing. In the RELAY trial, the ramucirumab is given every 2 weeks, which is different from the North East Japan study in which the ramucirumab was every 3 weeks and erlotinib on the standard dose.

TKI Plus Antiangiogenic Combination Therapy

One of the issues with the Tata Memorial Centre study that was done in India is this: It’s been observed that patients don’t cross over to get all the subsequent therapies that they do in the United States. For example, the use of osimertinib is much lower. To what extent do you think the findings from those studies will be relevant in the United States or western Europe with the treatment? Do you think the principle should be the same no matter what?

That’s a good point, John. In western Europe and in the United States, we have a facility for drug access, and we may not have that kind of pressure that other colleagues in other latitudes may have. In that approach, you’re questioning in 2 senses. In those countries, when they don’t have the resources we have, it is important. It is important because in the study you just mentioned, at least 40% of the people did not receive second-line chemotherapy. By moving the chemotherapy to the up-front setting, you’re giving everything together and getting that higher efficacy. The study shows that we have the patient use longer. That is 1 point. This is important.

The other thing we mentioned is about drug access. That’s a major issue but not for us. We cannot forget, John, that both studies were strongly positive for overall survival. Here the United States and in western Europe, we need to keep an eye on that. The studies are ongoing because we cannot hide the impressive data that came from those phase 3 trials.

You’re right. Personally, I’ve got mixed feelings with it. There was a study in the United States. It was a CALGB [Cancer and Leukemia Group B] study, and Pasi Jänne was the lead of this. It combined erlotinib with carboplatin and paclitaxel vs erlotinib alone. There was a bit of benefit; the benefit was mainly for the L858Rs, not to the exon 19, but it wasn’t a significant benefit overall. Here, the benefits were striking. They were undeniable. It’s overall survival. It’s clear-cut, as you said, in these 2 studies, and they involved gefitinib. Is there a difference of gefitinib vs erlotinib, or is there a difference with the populations? It’s hard to know, but I share with you the eagerness to see the osimertinib study. It will be very interesting to see how that comes out.

Let me play this out. Let’s say a study like that turns out to be positive. Knowing what you know right now, if you were treating patients in those places, would you use the chemotherapy and TKI [tyrosine kinase inhibitor] in the up-front setting, or would you pick a particular population based on what you see from this?

If I am in those areas, living there and practicing there, I would use the combination.

Yes, no question. I would use the combination. When you go deep in the data from the paper, and you say, “What was the problem?” You see carboplatin and pemetrexed. There was chemotherapy-induced myelosuppression and nephrotoxicity, but it’s only 4 cycles of carboplatin-pemetrexed. When I analyzed the dosing, the standard of care is this: After 4 cycles, you do the carboplatin, so you may expect myelotoxicity. But I don’t know, John. I’m sure that, in our practice [Florida Precision Oncology], it is rare to see myelotoxicity from pemetrexed alone or in combination with gefitinib when you give prophylaxis to the patient with adequate B12 replacement and folic acid. I don’t recall a patient of mine in the last 10 years for whom I have had that problem with myelotoxicity with pemetrexed.

No, they tend to do well. That’s right. We’ll see how the osimertinib studies work out, but it is an interesting finding. I’ll say that there’s 1 point that it illustrates for me. All of us, when we think about our own practices, think that everybody is going to get second-line or third-line treatment, but the fact of the matter is this: Things happen to lung cancer patients. Patients sometimes have unexpected problems that arise, or they get blood clots or infections and so forth. There have been a number of studies that show giving your best therapies or your best combinations up front is often a better approach than holding things for later because you sometimes don’t make it to later, even if you think you’re going to.

Absolutely. John, I also want to mention that, when the time comes, and we need to compare those trials, it will be difficult because what the trial from India includes was, for me, very impressive. They have more than 20% of the patients PS [performance status] 2, which is a very bad performance status. When we review the trials using TKIs, all of them have ECOG [performance status] equal to 0 to 1, so that’s something to keep in mind. Both studies also include those rare mutations that you just mentioned. They include 719, L61, and 768, which were not in the FLAURA study. They were not in the ARCHER 1050 study, and they were not on the afatinib trial initially because they removed them from analysis as you remember. Those 2 studies were what we see in reality, so that’s something to keep in mind.

I agree with you. It’s exciting that we have a good first-line option, but there’s clearly room to improve on it.

TKI Combination Clinical Trials: Practical Implications in US and UK

Eddie, let’s move on to talk about ways that EGFR inhibitors can be combined. As you noted before with the IPASS study, and we’ve had a number of other studies since then like AURA3, the TKI [tyrosine kinase inhibitor] is better than chemotherapy, but there is, of course, a question: Are there benefits to putting the 2 together and giving a combination of chemotherapy plus a tyrosine kinase inhibitor? Talk about your experience with that and studies that you think might be relevant to this.

John, there are 2 studies recently presented in our International Lung Cancer congress. I will go through them, in which TKIs were combined with chemotherapy. We need to understand that, in the past, we tried this approach, but it was an unselected population. It was everyone, and we couldn’t see any difference by combining TKIs with chemotherapy in those trials developed 10 years ago. With that, as you mentioned, we learned more about EGFR, we learned more about the mechanism of resistance, and we started to do exactly what you beautifully described: the sequence of first-generation, second-generation, and third-generation TKIs, and that toxicity has been lowered while efficacy has been moving up. Besides that, when we see those trials, I remember the trial and those in which we used chemotherapy plus TKI. There was another time when we used a combination of erlotinib with gefitinib. There was no difference.

In chemotherapy, we have also made a lot of progress. For the patient with EGFR, they go with a histology in lung cancer that we call adenocarcinoma. For adenocarcinoma, we have this chemotherapy called pemetrexed. The pemetrexed is, in my opinion, also a great development in oncology, specifically for thoracic oncology. If we give the appropriate prophylaxis with folic acid and B12, those 2 vitamins basically shut down the hematologic toxicity from this chemotherapy agent. We have also made advances in that regard.

There are 2 studies in which patients who were 

-mutation sensitive were randomized between gefitinib in combination with chemotherapy vs gefitinib alone. The idea is this: By adding chemotherapy to gefitinib, the efficacy will be superior because we know that at this moment, everyone is focused on monotherapy—monotherapy TKI, gefitinib alone, erlotinib alone, afatinib alone, or osimertinib alone. Can we do better? Can we improve upon what we have already obtained with those monotherapy TKIs?

There are 2 studies, as I mentioned before, and those studies have a bit of a difference between them. One was from Dr Yukio Hosomi from the North East Japan Study Group, the NEJ009, and the other is a study from a single institution in India by Dr Vanita Noronha. In those 2 studies, the primary end point was progression-free survival as with all the studies with TKI as well as overall survival. In those 2 studies, which were published this year, the progression-free survival was doubled for the combination of chemotherapy plus gefitinib vs gefitinib alone. The overall survival, which has now become the standard of care for us in lung cancer management as a first-line end point, was also positive in both studies.

Now the question became this: Will we need to do other clinical trials to evaluate that? In this particular study from North East Japan combining gefitinib, carboplatin, and pemetrexed vs gefitinib alone, they have an overall survival, John, of 50 months, which is the highest overall survival that we have seen. The question is this: How will that combination behave against osimertinib, for example? Perhaps osimertinib in combination with chemotherapy in the future can give us better responses. We don’t know. Both of these, which we will talk about later, are coming, but we need to wait for that.

We need to understand the concept of combining or adding chemotherapy to gefitinib in this case. It may make sense because what happens is this: As I mentioned before, the tumor is heterogeneous. You have a lot of subclonal population because not everyone has the 

 exon 19 and not all the cells have the L858R; they are the subclonal population. We are targeting those, but what about the others that are either less tolerant to the drug or those who will not respond to gefitinib? By making that pressure with those targeted agents like gefitinib, dacomitinib, and all those TKIs, we help those other clonal populations take over. That is where the problems come, as you mentioned before: all those mechanisms of resistance. Then what is there to do? By adding chemotherapy, we are covering that base, and it needs to be like that.

EGFR TKI Plus Chemotherapy Combination Strategy

You did a beautiful review of the first-generation, second-generation, and third-generation agents. You remember when afatinib, a second-generation drug, was developing. Afatinib was also studied in a real-world study called GioTag, and from that we may get some information that may be useful in clinic. Can you take us through your take-home message from the GioTag study using afatinib in the real world?

For afatinib in the real world, there are a couple of key points. Afatinib is clearly active. It has some CNS [central nervous system] activity. The CNS activity does not tend to be quite as high as osimertinib, but it does have CNS activity. Afatinib also has activity against atypical mutations. It’s the only TKI [tyrosine kinase inhibitor] that’s approved for 3 atypical mutations: the S768I, the 719 mutation, and the L861 mutations. Afatinib is a drug that has a broader spectrum of activity than osimertinib does, and it has CNS activity. In common use, people will often dose reduce it because at 40 mg, the starting dose, it is often found that patients have a lot of acne and diarrhea, and they often end up tolerating 30 mg better. I start everybody at the full dose and then can go to the lower dose if they need to, but I try to keep people at the higher dose and work them through for at least a couple of months before I dose reduce, unless they really need it.

It is worth saying that afatinib does have a place. It does have some activities that the others don’t. You know the approvals, and the activity has been better than the first-generation drugs in general. It does still have a place in the real world as an EGFR inhibitor.

You mentioned something important. This is important for the audience and our colleagues who listen to you and me today: the issue with afatinib. When I usually go around and talk to colleagues, they sometimes believe that starting afatinib at a lower dose of 30 mg, for example, is the best way to do it because there was also a study, a retrospective analysis, that showed that that dose, the efficacy has no drop. I feel that the intensity of the treatment, which you described well, is the way to go. Try to control the disease, try to eliminate as much as we can, because at the end of the day, we know that resistance is coming. By starting on a lower dose, there is a gap that means that the resistance may go quicker than by starting at a higher dose.

It’s an important point. It’s key for the EGFR inhibitors in my case, and we [at The University of Texas MD Anderson Cancer Center] have done a lot of testing of other drugs that also have a lot of skin toxicity and diarrhea, like poziotinib as well, where we need to manage it carefully. It is manageable, but it needs to be managed carefully. The key is education of the patient up front and setting expectations. The way I always explain it to patients is the following. I explain how, for these EGFR inhibitors like afatinib, dacomitinib, and erlotinib, you are going to develop a rash and acne. I sometimes have pictures to show them so they know what’s coming. Personally, I give them clindamycin gel, and I give them an antibiotic like doxycycline or minocycline up front. I tell them to buy some topical hydrocortisone as well and to buy Imodium [loperamide hydrochloride], so that they have all that in their house and they’re not wondering in the middle of the night what they’re going to do.

I tell them to expect that it’s going to get worse for the 3 to 4 weeks, but then it’s going to eventually level off and start getting better. The adverse effects won’t go away by 2 months, but they will get better. If they understand that it’s going to get worse and then it’s going to slowly get better and level off, then they’re not so quick to dose reduce at the first sign of trouble. For doctors, if you set that expectation, patients are willing to tolerate it if they know it’s going to be a month. If they’re afraid it’s going to be like this for years, then anybody would want a dose reduction. That’s why I try to get patients through 2 months of therapy. I say, “Listen, it’s going to get worse, then it’s going to get a little better. Let’s try to get to 2 months before we start thinking about dose reducing” to try to make sure patients get that full dose.

Data Updates From the GioTag Study

One of the problems that is critical for these 

 patients is, as you know, 20% to 30% of patients with this disease present with brain metastases. Brain metastases is something that could be the Achilles tendon for these patients. Can you tell us more about that? For example, if a patient has a brain metastases, asymptomatic or with symptoms, would you give a chance to the patient to respond to osimertinib? As you know, that is responding quickly, or do you go with stereotactic radiosurgery, like CyberKnife? Tell us about your management of early brain metastases asymptomatic patients.

This is a great question. We have a lot of debates about this. I’m sure you do too, Eddie. In fact, we had 1 this week. What’s commonly happening is that somebody presents with 1 or 2 small brain metastases. My approach, because osimertinib has such good CNS [central nervous system] penetration, is if they’re small and asymptomatic, I offer the patient the option of not getting radiation up front and just using osimertinib. It’s interesting that some patients very much like this because they want to avoid radiation. Some are concerned and want the security of getting the radiation first.

The situation we’ve had recently—and I’m sure you’ve run into this situation as well—is if we’re staging a patient, and we don’t know if they have 

 mutation, and you find that there’s brain metastases. In this case, we referred the patient to our neuro-oncology group, and they’re a wonderful group. They’re great at giving gamma knife. They can get it ready to go, and the result comes back and says “

 mutation.” In 1 case, we literally had the patient going into the room. Then the question is this: Do you pull the patient out of the room and say, “Listen, let’s wait a couple of months and see if you need it?” Or do you go ahead with the radiation up front?

 result back before I do radiation because you can avoid radiation in a lot of patients or avoid it for a long time by just starting with osimertinib. The responses are good.

There are a couple of other issues that come up. One example we had with the patient was this: They were all set up for radiation, we got the result back, and they said, “I’ll just go ahead with the radiation.” Sure enough, that ended up being a patient who developed radiation necrosis later, and we spent months dealing with radiation necrosis. Even though the gamma knife worked beautifully, it’s not a free lunch. It’s much better tolerated than whole-brain radiation, but you do get radiation necrosis and other toxicities from it. This patient has been dealing with radiation necrosis for months now as a result, so I wish we had stopped and given the osimertinib there as well.

The other situation is this: What if they’re symptomatic? I’ve had some cases where somebody may be symptomatic, but they may not be. If you ask them, “Have you had headaches lately?” They may respond, “Yeah, I have had headaches. Maybe I’ve had a bit of double vision now that you mention it, but it hadn’t been bothering me. I thought I just needed new glasses.” This is often the situation if somebody is mildly symptomatic. Do you give the TKI a chance? I would often consider it. On the other hand, at least my approach is this: If I see that somebody has gotten serious brain metastases with a lot of edema, if it’s causing dilation of their ventricles, or if it’s near their brain stem or something else and I feel like we’ve got very little margin to wait, I tend be safer and go ahead with the radiation right away while I’m doing the tyrosine kinase inhibitor. That’s the approach I’ve taken, but osimertinib clearly does have good CNS activity.

Another issue to mention is what dose to use. There was a study called BLOOM that tested a higher dose instead of the standard 80-mg dose. They went to 160 mg of osimertinib. For me, that’s something I consider with my patients with brain metastases: going to the higher dose, just because you’ll get better levels in the CNS of the drug as well. Even though 80 mg is the approved dose, I at least consider the higher dose if I’m concerned that somebody has got serious brain metastases. This is an area we still struggle with all the time. What’s your approach if somebody is asymptomatic up front with new brain metastases?

As you said, there are symptoms that tell you that we better move with CyberKnife depending on the lesions and the locations. That is critical. Another area that is a concern for me is the cerebellum because that is a critical area. It has vital functions for a human being located in the cerebellum, and it is easy for the patient to get into trouble. If there are cerebellum lesions, I may go quickly with a CyberKnife or gamma knife depending on how many lesions they have.

I agree with you. It depends on what kind of symptoms the patient describes. With so many patients we see in clinic with lung cancer, I still haven’t seen any case in which I need to use, for example, osimertinib double dose because of leptomeningeal disease, for example. I have not had that case, thank God, because you know those are difficult cases to treat. I didn’t have the chance to use the 160-mg dose, but that would be my approach. If the patient has symptoms, go deep on the history prior to jumping immediately to the radiation therapy. I totally agree in that sense with you also.

The Use of TKIs in Patients With Brain Metastases

How has the development of tyrosine kinase inhibitors [TKIs] been from the third-generation drugs, and how have we ended up today with a third-generation osimertinib? You can take us to that pathway in the last 10, 12 years.

It’s hard to believe. It’s even longer than that that we’ve been testing these drugs. I treated my first patient with gefitinib in 1999, believe it or not, when we had those early trials going on, and that preceded the discovery of 

 mutations. Of course, the first drugs that were in the clinic for EGFR were gefitinib and then erlotinib, so we think of those as the first-generation drugs. They’re certainly active.

The hallmarks of the first-generation drugs are the following. First, they have toxicities related to inhibiting wild-type 

. Those toxicities are familiar to any of us who treat patients with EGFR inhibitors. This is the acneiform rash and commonly some diarrhea. They often get fingernail issues like paronychia, and then some other rare things, such as adverse effects in the hair and in the cornea mucous membranes and other rarer adverse effects as well. This all has to do with the fact that wild-type 

 is present in hair follicles; it’s present in a lot of other tissues. Here’s an interesting tribute point: 

 was initially discovered by Stanley Cohen. In fact, he won a Nobel Prize for this because it was found to be a factor that was critical for teeth erupting in mice. It was key in the differentiation of certain epithelial surfaces. In mice, when you knock this gene out, they get a funny phenotype with their hair; it’s called a wavy phenotype. That wavy phenotype has been known for a long time. Mice don’t get the acne; they get other skin and hair effects.

Those were the first-generation drugs, and resistance emerges. Typically, these drugs will give a PFS [progression-free survival] on the order of 9 to 12 months, and the most common alteration-associated resistance for these is the T790M. They call this the gatekeeper mutation, and this prevents the binding of the first-generation drugs. The first-generation drugs are irreversible inhibitors, so if they get a gatekeeper mutation, they can’t bind avidly enough into that pocket, and that gatekeeper mutation prevents them from inhibiting the receptor.

The second-generation drugs are dacomitinib and afatinib. They’re a bit more potent, and they can bind irreversibly, but T790M still often emerges, and they still have a lot of wild-type activity. Both afatinib and dacomitinib have a lot of rash, and they have a lot of diarrhea associated with them. It’s manageable in the setting of the clinical studies. If you look at outcomes for the second-generation agents vs. the first-generation agents, they’re a bit better. Afatinib has been tested against gefitinib head-to-head; it was a bit better. Dacomitinib was tested against the first-generation drugs, which was the ARCHER1050 study, and it prolonged overall survival in that study. The median PFS was about 15 months, so that was a real difference, but it came at the cost of a fair amount of skin toxicity. It was manageable but not trivial.

This brings us to the third-generation drugs. The big differences with the third-generation drugs are 2-fold. The first is that they bind irreversibly, so they have a covalent linkage. The details here are that they form a covalent linkage with the cysteine near the binding pocket: It’s called the C797. The reason that’s important is because, if that cysteine gets mutated, then third-generation drugs don’t work anymore, so 1 mechanism of resistance to third-generation drugs is the C797S mutation.

The advantage of third-generation drugs, of which osimertinib is the only 1 that we have approved right now, is that they have a much better specificity. They inhibit the mutants much better than they inhibit the wild-type receptor, so they have much less skin toxicity and less diarrhea, and they’re better tolerated overall. Osimertinib is a well-tolerated drug. It’s more active, and it will inhibit the T790M mutation, so you don’t get the same mechanism of resistance.

Interestingly, we [at the University of Texas MD Anderson Cancer Center] and other groups have recently reported about how you get resistance to third-generation drugs. You often get bypass mechanisms. You get a different kinase that’s active like MET amplification or 

 fusions, and there are a number of others that can happen. 

 [human epidermal growth factor receptor 2] can get activated, and so forth, in which case, you may have to switch therapies. There are other mutations that can emerge. The important study that was done was called the FLAURA study, and that compared osimertinib with the first- and second-generation drugs: erlotinib and gefitinib. It was substantially better for PFS, with about 18.9 months vs. about 10 months, so there was a big difference. As we said, most of that benefit was seen in the exon 19 deletions, not in the L858R, and it prolonged overall survival as well. Third-generation drugs are clearly an advance with better tolerability.

Another factor I didn’t mention is that there was better CNS [central nervous system] penetration for osimertinib. There was much more activity in the CNS. You don’t get T790M as a mechanism of resistance, but you do get other mechanisms of resistance. One of the challenges that emerges is if you get resistance to osimertinib. What other targeted agents do you go to? We don’t have any other approved agents if you get osimertinib resistance. In some people’s minds, they consider, “Maybe I should start with a first-generation drug or a first-generation drug in combination,” like the ramucirumab studies we talked about earlier. If you get resistance, you can use osimertinib and keep that in your pocket for later. There are pros and cons for those 2 approaches.

That’s how the field has evolved from first-generation to second-generation and now to third-generation drugs. It’s been a clear advance with clear progress as we’ve gone forward.

Approved Frontline TKIs for EGFR-Mutated Disease

Here’s another issue that’s emerging. You know the ADAURA results just presented at ASCO [American Society of Clinical Oncology Annual Meeting] said that there looks like there is a good benefit for using EGFR inhibitors in the adjuvant setting if somebody has an 

 mutation. It’s not FDA approved. We’re eagerly following those results and that submission, but it raises the question: Do you send profiling in the earlier-stage patients? Do you send it early in the work-up when you’re not sure if they’re metastatic? Or do you wait until you’re sure they’re metastatic before you send it?

That’s a good question, John. I have been using it, and I’m pretty sure you have been as well for many years. We have been using next-generation sequencing for metastatic non–small cell lung cancer. Now I need to sit down and think, “What vendor can help me do a PCR [polymerase chain reaction] for 

?” I have to think about it, because when you look into that, basically everyone has a panel, and a minimal panel is with maybe 25 genes, 50 genes. It can go up to 600 genes, so at this moment if I have 2 patients—to answer your question—who are postsurgical adenocarcinoma, 1 is a never smoker and the other was a former smoker, I send the panel but try to do a smaller panel.

Right. Let me ask you about something else. There have been a number of studies: studies with erlotinib—for example, the old CALGB study—and studies with afatinib, that have shown that L858R seemed to do worse than exon 19 deletion in terms of outcome. They both respond to drugs. Is that consistent with your experience? Do you shape your therapy based on that at all right now? How are you taking that into account?

Yes, I agree with those observations. Since the IPASS trial in which gefitinib was compared with carboplatin–Taxol [paclitaxel] more than 10 years ago by Tony Mok, we know that those 2 common and sensitive 

 mutations—exon 19 and L858R, which you said—are not equal. They are not equal, and they are different in how long the patient will remain on therapy, which is something that we call progression-free survival. When you see all the data from that time until now, and there is either a post hoc analysis, a subset analysis, or a prespecified analysis, there is always a difference between exon 19 and exon 21, which is the L858R. Although they will respond well, how long the patient will be without progression is always lower for the L858R. It doesn’t matter what TKI [tyrosine kinase inhibitor] you use: You will always see the difference. For that observation, I agreed.

I haven’t done the difference in terms of therapy, and we’ll probably discuss that later. Until recently, there is a clinical trial that we will also discuss later: the RELAY trial, in which there is a difference there compared with what we have seen in the past through several clinical trials.

Right. It’s interesting that this held up with the osimertinib study. It was the frontline study, FLAURA, where osimertinib was compared with gefitinib or erlotinib. The benefit for osimertinib was striking in the exon 19 deletions, but it was much more modest in the L858R mutations.

Correct, exactly. Even with osimertinib—which as you know is a secondary drug by the NCCN [National Comprehensive Cancer Network] and is also the preferred agent from the NCCN, the third-generation TKI, which is very powerful with a good toxicity profile—there was a difference between those 2 mutations.

Molecular Testing Prior to Metastases

We have liquid biopsies that are available. When and where do you use the liquid biopsies? Will you do it up front for patients, or do you do it if there’s not enough tissue? How are you using it? Is there any resistance?

 For the liquid biopsy, John, the technology came to the United States at the end of 2014. I started to use it at the same time. I was sending tissue and liquid biopsy simultaneously for many reasons. At that time in 2014, and still in some places, there was still this rule for Medicare patients: If they get a biopsy at a hospital, that hospital cannot send the biopsy until after 14 days after the procedure. There’s a lot of time that we are losing, and the patient with stage IV lung cancer is waiting for us to know. That’s why I decided at that time to send a liquid biopsy because, on the first visit of the patient to your clinic, you’re able to send that blood test, and you may have an answer quickly. Now, I remember when I sent my first liquid biopsy: It was about 2 weeks. Now you can get an answer from 3 to 7 days, and you may have an entire profile that will help you start therapy right away if you find a specific driver mutation. The answer for you is that I’ve been using that at the same time.

As time passed, a colleague of ours from a different institution also proved in a prospective trial that both of those technologies, using NGS [next-generation sequencing] in the tissue and NGS in the blood, are complementary and are more exclusive. Perhaps some of them missed the driver mutation that we are looking for because we need to remember that, No. 1, no technology is perfect. There is tumor heterogeneity depending on where you do the biopsy, and it is also so in the liquid biopsy: the same thing happens. A liquid biopsy is great if you have the DNA of the tumor circulating in the blood, but if the tumor is not shedding tumor cells to the circulation, you will not find the mutation regardless of whether you have the best technology. None of those technologies or labs, either tissue or plasma, is perfect. Both work together.

Yeah, I agree with you. The study, there was 1 that you mentioned from UPenn [the University of Pennsylvania], in which they showed that, if you send both, you find out that tumor biopsies alone miss something like 10% of cases. If you think you’re covering everything with tumors, you’re not: There’s some you could be missing. There is a good argument for sending both of them, and you’ll be getting it quicker. I can say that, at The University of Texas MD Anderson Cancer Center, we’re a big cancer center, but it still takes a long time between ordering the biopsy and getting the patient scheduled. They then get the biopsy, and sometimes, the tissue is not sufficient, and you can waste a few weeks on that.

With plasma, at least it is in our hands. We’re able to detect drivers in something like 80% to 90% of the cases if they’re there.

Yeah. Another thing that we haven’t mentioned, John, is the concordance between the 2: It’s like 98% concordance. If the problem is in the blood, you know that it will be in the tissue, so this is important.

Yeah. That’s been our experience as well. We recently reported this: If it’s a positive test by plasma, you can have a high degree of confidence because the positive predictive value of the blood test is extremely high, but if the blood doesn’t come up with anything and your suspicion is still high, let’s say it’s in a nonsmoker or a younger patient, it’s probably important to wait for the tumor tissue as well.

Metastatic NSCLC: The Role of Liquid Biopsy

Hi, I’m John Heymach. I’m the chair of thoracic head and neck medical oncology at The University of Texas MD Anderson Cancer Center in Houston, Texas. During this 

®program, we’ll be having a conversation about 

-mutated non–small cell lung cancer. The conversation today will be with Dr Edgardo Santos. Dr Santos, would you like to introduce yourself?

Thank you, John. It’s always a pleasure to be with you. My name is Edgardo Santos with Florida Precision Oncology, precision thoracic oncology, in Aventura, Florida. I am the medical director of research services.

Wonderful. I’ll let the audience know that Eddie and I share a number of patients, so these are real-life issues that we get to discuss on a day-to-day basis. Let me start, Eddie, and let’s talk about molecular testing in non–small cell lung cancer. This is such an exciting and rapidly changing area because now, just in the last couple of years, the number of new targeted agents we have is exploding. We now have at least 7 targetable oncogenes for which we routinely need to test, so this is something that is changing month by month with these new approvals. Let me start by asking this: What’s your approach for molecular testing in patients, and what platform or approach are you using? Are you still including things like FISH [fluorescence in situ hybridization]–based assays, IHC [immunohistochemistry], or NGS [next-generation sequencing]? What are you doing for 

As you know, this field has changed significantly in the last few years. We are more ahead compared with the initial times when we were using immunohistochemistry and perhaps PCR [polymerase chain reaction] and FISH analysis. The technology has advanced so much nowadays that most oncologists are using next-generation sequencing, which allows us to cover multiple gene abnormalities. A big part of that genome is that you want to say that, by doing next-generation sequencing, we may have a better understanding of the spectrum of what’s going on with a patient. It also uses a small piece of the tumor tissue, which is valuable for the future of management and research. It’s always better not to utilize all the entire block where the tissue from the biopsy is.

 is perhaps 1 of the most common driver mutations that we have, specifically in lung cancer. You start with the panel. If you order next-generation sequencing with any kind of vendor, most likely the 

 is there plus many others, so that’s what I use. I don’t recall the last time I sent a specific patient for just 

The other thing that’s new, as you just mentioned, is the advances we have done on next-generation sequencing now using the blood: It’s what we call liquid biopsy. They also use next-generation sequencing as the platform, and that has helped us make a faster, more liquid diagnosis because the results are coming faster than the tissue molecular profile would. That has impacted the patient care, not only in lung cancer but also in other tumor types.

. You and I have both been treating lung cancer for a long time. You remember early on there were a lot of data about 

 amplification and doing a FISH for that. Are you ever doing IHC or looking for 

 amplification, or are you focused on the mutations?

, I focus on the mutation. I’m not using an amplification as you mentioned. There are some cases for which immunohistochemistry was also used. I focus on the mutation. As you know, there are several 

 mutations, so that’s what I do. What do you do in your practice?

It’s the same. We used to do IHC, and we used to do amplification, but now we’re focused on those mutations. It’s worth discussing. We used to have a PCR [polymerase chain reaction]–based platform that focused on the 2 classical mutations: the L858R and the exon 19 deletion. With the next-generation sequencing, we’re covering a much wider range of potential mutations. We have T790M, and we’re getting the exon 20 insertions and the other atypical mutations. What’s your approach there? Are you covering a wide range of mutations, or is your platform catching the classical mutations only?

 exon. We cover all of them that you just mentioned: those atypical or uncommon mutations, some of which are sensitive. We cover all of the 

Yeah, it’s great. It is not just important for selecting standard therapies. We know our standard therapies for 

-mutant lung cancer after the classical mutations, but we do have afatinib now approved for some atypical mutations: the L861Q mutation, the S768I mutation, and the 719 mutation as well. You want to be profiling them. From a research perspective, this is an area that we’re interested in at MD Anderson. We now have some drugs that are active against exon 20 insertions as well. Even though they’re not FDA approved, it’s important that we identify them so we can get those patients on clinical trials. Has that been your experience as well?

Yes, John, it’s the same thing. Either we put a patient on a clinical trial, or we at least know the aggressiveness of the disease and look for a therapy if there is no research available. I completely agree with you. For exon 20 insertions, as you know, there are some drugs in latest development stages, and hopefully we’ll have them soon.

Here, I’ll mention that we take the same next-generation sequencing approach that you mentioned as well, because if you think about all the different alterations besides 

, we’ve got to cover it. Before, we just had 

 fusions, for which RET inhibitors are approved. We’ve got MET exon 14, with the approval of capmatinib as well, and of course 

 fusions. We used to do a lot of these FISH assays for the fusions, but for the reason you mentioned before, it seems impossible to do all those FISH assays. That’s 5 different FISH assays.

Are you all doing FISH assays or are you trying to cover it all with the NGS?

As you know, we talked about next-generation sequencing, but that technology has also improved because now several vendors also have RNA generation sequencing. That next-generation sequencing is targeted and developed to the DNA analysis, and now RNA is also on board. It’s difficult to miss a fusion protein or an abnormality when you combine the NGS of both DNA and RNA. We will be able to catch a lot of those things that you just mentioned: all those official proteins for which FISH is good, like RET and NTRK1. Sometimes they’re very picky, but with the RNA on board, we’ll be able to catch them. It’s good because we look forward to seeing an increase in the number of patients being diagnosed with them because we now have targeted therapy for those patients.

Right. The therapies for those rarer alterations, the things like the fusions, are really good. The patients can have a dramatically improved outcome, so there’s probably nothing more important an oncologist can do than to make sure they get the right driver mutation for patients.

Edgardo S. Santos Castillero, MD, FACP, Florida Precision Oncology

Articles