'Mind-boggling' Immunotherapy Results, Plus New Combos, Targets for NSCLC: ASCO 2019

H. Jack West, MD; Charu Aggarwal, MD; Sandip P. Patel, MD


June 07, 2019

This transcript has been edited for clarity.

H. Jack West, MD: Hello. I'm Dr Jack West, associate clinical professor and the executive director of employer services at City of Hope Comprehensive Cancer Center in Duarte, California. Welcome to Medscape Oncology Insights. Joining me today are thoracic oncology experts Charu Aggarwal, the Leslye M. Heisler Assistant Professor for Lung Cancer Excellence at the University of Pennsylvania; and Sandip Patel, an associate professor at University of California, San Diego. Thanks so much for being here.

Sandip P. Patel, MD: Nice to be here. Thanks for having us.

Survival Increases in Advanced NSCLC

West: Today we're going to talk about some of the new lung cancer data presented here [at the American Society of Clinical Oncology meeting]. One of the most interesting things I would say is actually not new data but the revisiting of an older trial. Dr Eddie Garon from University of California, Los Angeles, reported on more than 500 patients with advanced non–small cell lung cancer (NSCLC), most of whom were previously treated, who received pembrolizumab in a very expanded phase 1 trial.[1] We saw a 5-year survival in these patients that was about 30% in the patients with high PD-L1, and about 25% overall. I think this is remarkable on all sorts of levels, one being that we're even talking about 5-year survival in advanced NSCLC when a few years ago we didn't even report 2-year survival consistently. Now we even have a relatively unselected population who can have a realistic potential of living years and years. Of course, in a more selected population that we are honing based on PD-L1, it's approaching 1 in 3. What were your impressions of this?

Charu Aggarwal, MD: We were one of the participants in the trial and this was actually one of the first trials that I started when I was a newly minted attending. I'm happy to say that some of the patients I first met in my attending clinic subsequently went on to this trial. There were patients with brain metastases and patients destined to die within the next year or two, and they are still alive in follow-up. It's going from single-digit 5-year survival in NSCLC in these patients without actionable mutations to a point where we're seeing a quarter of these patients alive 5 years and beyond. It is breathtaking and mind-boggling. I'm just so impressed at how far we've come. We can solidly now say to our patients, "Here is a very viable treatment approach for you and you [may] live a long time."

West: We've had so many advances in narrow subsets, often with some of the same features. If you're a never-smoker with an adenocarcinoma, there are a lot of targets and a lot of targeted therapies. But what is especially remarkable about the pembrolizumab data is that this is in a group of people who we really didn't have a lot to offer before. We finally have something meaningful to talk to them about, offer them, and give them hope. Sandip, can we realistically talk about or think about patients being cured? And related to that, what has been your practice in terms of keeping people on this or not? In your mind, is this something that is suppressive that you need to be on longitudinally, or can we potentially use the immune system to eradicate the last cancer cell and surveil after that?

Patel: It's absolutely a key question because these patients are going to live a very long time. One of the key salient features of immune checkpoint blockade is that your own immune system has kind of figured out the cancer; you've got to reset the clock by giving these immune checkpoint blockade agents such as pembrolizumab. I do think some of these patients may actually be cured and be off therapy and be in remission. In my practice, it depends on the incidence of immune-related adverse events (IRAEs), the autoimmune side effects that we get from immunotherapy.

A recent paper from Rich Pazdur's group at the US Food and Drug Administration[2] showed that it's a bit of a double-edged sword; sometimes the patients with the most elite responses often have the highest IRAE rate. In patients who are having difficulty with some of the more chronic IRAEs, I'll consider trying to get them to 2 years as much as safely possible. Maybe I'll get a PET scan and try to incorporate some of what we've learned in the oligoconsolidative setting with stereotactic body radiation therapy (SBRT). Often these patients who are responding have a few lesions and they may be eligible for Gomez/Palma-type approaches[3,4] and have a treatment-free interval, which ideally may be for the rest of their life. That is the promise of immunotherapy. These drugs technically are prodrugs; the actual drug is the T cell within your body. Those stay in with you, and we just have to make sure they don't fall back asleep. How to personalize that duration is an ongoing question. I do think we have some cures here, and it's great to see the 5-year data. Maybe in 5 years' time we'll have the 10-year data.

West: The fact that we're even asking whether we're curing or not is a great question to have.

Aggarwal: Eddie Garon put it beautifully yesterday when he was asked this question and he said, "We don't know." I would love to say that these patients are cured but they are potentially cured. Like you said, we need to wait for the 10-year data. Five years ago we could not even say that we wanted to see 10-year survival data with these metastatic NSCLC trials. These are very exciting times.

EGFR Mutation

West: Let's turn to patients with driver mutations. We've had some new advances in the epidermal growth factor receptor (EGFR) mutation group, the largest group we've been studying, for more than a decade. Many of us have kind of consolidated our approach around first-line osimertinib but some legitimate trials challenge that concept. The RELAY trial[5] out of Japan looked at patients with an activating EGFR mutation getting first-line erlotinib with or without ramucirumab. Not only was there a significant improvement in progression-free survival (PFS) and what looked like potential for an overall survival (OS) benefit, but it was not accompanied by prohibitive toxicity and you preserved the potential to pursue osimertinib, at least in the subset that has a T790m mutation after, and chemotherapy. The median PFS was a little over 19 months (which is in the same ballpark as osimertinib), but this was in patients who didn't have brain metastases so it was not the exact same population. Does that concept of pursuing a more aggressive approach in the first-line setting and preserving the potential for further options down the line have any appeal to you these days or are you still largely set on osimertinib?

You never get a second chance make a first impression against aggressive cancer.

Aggarwal: The median PFS is encouraging in that it came in just over 19 months, but these were patients who did not have brain metastases at presentation. The FLAURA study[6] actually included patients with brain metastases. The question of whether we should come in with an approach like this first brings us back to conversations from 2-3 years ago when we used to question whether to start with erlotinib and then come in with osimertinib. We were waiting for the FLAURA data to sort of add up the math of the PFS and see what would be better. But osimertinib really blew it out of the water. We saw significant improvement in PFS. We saw activity in the brain. We saw tolerability. It won on all fronts. I strongly believe that antiangiogenesis has a place. I just don't know if we're ready to introduce an intravenous drug in patients in whom we could otherwise get away with using a very well-tolerated oral drug. There is an ongoing study at Memorial Sloan Kettering looking at osimertinib with bevacizumab. I am very interested in that because I do believe there is better brain penetration with osimertinib.

West: ECOG is doing a trial with this as well.

Aggarwal: It's encouraging data but I'm just not sure if it will change our standard of care just yet.

Combination Therapy

West: I would not so much say that it will or should as much as we should step back and acknowledge that there are other approaches that are legitimate considerations. Also, we eventually need to think about the whole longitudinal strategy over a course of multiple lines and years, because we don't know yet how the patients are going to do post-osimertinib. In my limited experience, fortunately a lot of the patients who started on it are still on it, but it can be challenging and patients may not do that well in the years afterwards. It's a complex question. Sandip, do you see a potential place for this? How important is control of the central nervous system (CNS)?

Patel: You never get a second chance make a first impression against aggressive cancer. If you look at all of these studies in which there are carefully selected patients who are more fit than the general population, the crossover rate is anywhere from 60% to 70%, and these were with effective targeted therapies. My general bias is that you want to use your best therapies upfront, and I believe that osimertinib is the best therapy upfront in the EGFR space. But if you can take that premise of using the best therapy upfront, what is the best therapeutic combination to make sure [we treat] those clones that may not respond to EGFR blockade? Is that with an angiogenesis inhibitor plus osimertinib? Is that with chemotherapy plus osimertinib? These are the kinds of trials we've seen in the refractory setting or in Asia upfront with some of the first-generation inhibitors. It's similar to immunotherapy. We typically did biomarker-selected anti-PD-1 monotherapy, and then KEYNOTE-189[7] and KEYNOTE-407[8] really showed that by doing chemotherapy plus immunotherapy, you can potentially have the best of both worlds. What, exactly, may be the best combination with osimertinib in the frontline setting remains to be seen.

West: This year we had a study by Dr Noronha and colleagues[9] out of India that looked at gefitinib with or without carboplatin and pemetrexed in the first-line setting for patients with an EGFR mutation. They were a little more liberal in a few ways: They had some patients with the less common noncanonical EGFR mutations, and they had about 20% of patients with performance status of 2. The results in both of the groups were a little less favorable than we might expect, but that trial showed the same findings that we saw out of the NEJ009 trial[10] from Japan last year, which really impressed a lot of us by showing not just an eye-popping improvement in PFS but a 14-month improvement in OS. Gefitinib is not an agent that we would generally favor, but I would say that the setting where the probability of attrition, of losing patients who drop off and never get the chance for beneficial chemotherapy, and also the idea of the potential benefit of hitting it hard right from the from the beginning are factors. We've seen in very different settings, like prostate cancer, that there can be a disproportionate benefit to giving the same treatment early on rather than later on, [whether that's due to] less tumor burden or performance status. Would you say that in 3 or 5 years we're likely to be doing combinations, or do you think it's still going to be a perfect storm of multiple choices and no clear judgment on what is a best approach?

Aggarwal: I think we'll be doing combinations. I just don't know which combination it would be. At this point, knowing that osimertinib is so well tolerated, it's a hard sell for me to tell a patient they need to have chemotherapy. Although, having said that, carboplatin and pemetrexed are relatively easy to administer. You could presumably think of a scenario where you could do carboplatin/pemetrexed plus osimertinib, which some of us were discussing that we actually already do in the post-progression setting to sort of maintain that CNS control and provide some response rate with the chemotherapy. So, I do think that in a few years we'll be doing combination approaches. Will it be bevacizumab? Will it be ramucirumab? Will it be chemotherapy? I think this remains to be seen.

West: Sandip, you mentioned that you have patients who are potentially longitudinally on immunotherapy. We all do. Some patients may judge that they are perfectly happy to do what they need to do to get the best long-term outcome. Your thoughts here?

Patel: It is very interesting. You need to have a dominant mechanism for a combination in the frontline space because you're basically preventing that. The reason osimertinib exists is because it takes care of T790m, which is a dominant mechanism of resistance to first-generation inhibitors. I'm not sure that there is going to necessarily be a dominant mechanism of resistance that we can give to all coming patients in the frontline setting. Chemotherapy plus osimertinib in the frontline space may actually have a niche because chemotherapy basically takes care of any fast-growing clones. I think the second-line space is where we will see all of these novel agents, such as the JNJ bispecific, the MET inhibitors from multiple companies, and the HER3 antibody-drug conjugate. In a very biomarker-specific fashion, [patients may] develop pathway resistance X phenotype, develop pathway Y, or develop pathway Z. That is where you may actually get these adaptive combinations, some based on unique resistance patterns. It's probably going to look a little different in the frontline setting, but if there is not a dominant mechanism, that is where chemotherapy may actually come into play with osimertinib. In the refractory setting, it does not look like there is a dominant mechanism yet—maybe MET amplification more than average, but not enough that you would necessarily want to give everyone a MET inhibitor upfront. So, in the second-line space, I think we're going to see a little more heterogeneity based on the patient's unique resistance phenotype.

'Elusive' Targets

West: Beyond EGFR, there were quite a few very promising findings about targeted therapies for what have previously been very elusive targets that we wondered whether we'd ever be able to address. Sandip, you were present for an early presentation on KRAS targeting which has been seemingly taunting us as a very common but extraordinarily elusive target. Can you speak to that?

Patel: KRAS and exon 20 insertion, whether it's EGFR or HER2, have kind of a similar problem in drug design. These are very hard binding pockets to hit for small-molecule inhibitors. We've done well at designing these, but KRAS and exon 20 insertions have been very difficult to target until recently. A phase 1 study from Amgen[11] for a KRAS G12C-specific inhibitor (so it's not a pan-KRAS inhibitor) has six patients with lung cancer with that specific perturbation. It's very early preliminary data, but two out of six patients had a response. As federal taxpayers we spend nine figures each year trying to trap KRAS. To get a little bit, at least on a specific subtype, with preliminary signs of efficacy for a gene that historically has been called "undruggable" I think is huge. Mirati[12] has one too. Exon 20 insertion drug design is very similar. Our traditional EGFR inhibitors do not target that binding pocket and have the same issues with drug design. The ability to have not one but two agents now that potentially target that biology is particularly provocative.

West: One of the other appealing yet frustrating things is that KRAS represents 20%-25% of patients with lung cancer. But we need to get more granular than that, and that may be one of the keys. You mentioned exon 20 EGFR mutations and we did see some interesting stuff with that. Charu, what were some of your impressions of other agents or other targets that you could see a clear path to having really effective treatment approaches for our patients with advanced NSCLC?

Aggarwal: The lung oral session today was in line with the theme of the meeting: "Caring for every patient, learning from every patient." Christine Lovly put it beautifully, saying that we actually owe it to our patients to conduct molecular genotyping on all of them. She highlighted that nonsmokers can have actionable mutations. For example, RET fusion can be detected in one third of smokers. We should look for it. There was very interesting data on a drug called BLU-667. Last year we saw results from the phase 1 study[13] that showed significant responses across all solid tumor types, and today we heard expansion cohort data[14] from lung cancer. It was so impressive to see that irrespective of brain metastases or prior chemotherapy or immunotherapy, patients did really well with this drug. They had significant responses. But we will not be able to offer these patients this therapy if we don't look for it. What was really remarkable was that if you pick up the fusion in liquid biopsy, you can still afford that benefit. I've been a huge proponent of testing liquid biopsies and I think we owe it to our patients to be able to afford that significant benefit. And we saw not one but two drugs with activity against MET exon 14 skipping mutations.[15,16] It's so humbling to see that now we're at a point where we're actually dissecting the resistance mechanisms to MET. It's amazing. It's just mind-boggling that we're able to really fine-tune what a patient needs. We're at the cusp of precision oncology and we really need to hone it.

West: That, and the more the more you look, the more you find. These are things where, if you're not doing the testing, you're never going to identify these patients. Charu and Sandip, thank you so much for joining me for this great discussion. Thank you, the audience, for joining us. This is Jack West for Medscape.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: