Treatment Algorithms and Resistance Mechanisms in NSCLC

Neal Navani, MRCP, MSc, PhD: Gary, what about resistance mechanisms to osimertinib? Christian touched on them earlier. What do we know about some of the resistance mechanisms to osimertinib? Can you tell us a bit about that?

Gary Doherty, MRCP, PhD: There are many pathways that are involved in osimertinib resistance. We have innate resistance. Not every single patient responds—presumably because of the presence of co-mutations that are already present in the nascent tumor—but for patients who progress on osimertinib, there are very clearly defined resistance pathways. For about 40% to 50% of patients, we do not know what the resistance pathway looks like. It will be interesting to look at epigenetic changes within these tumors to see if we can figure out why some patients resist therapy, despite lacking a clear mutation or a genomic alteration. There are many pathways that have been identified. There is a slightly different spectrum of resistance mechanisms to osimertinib for patients with T790M-positive disease vs those treated with osimertinib in the first-line setting. For around 20% to 25% of patients, it involves either EGFR gene amplification—so that on-target mechanism—or an EGFR mutation, with the most common of these being the exon 20 mutation C797S; usually, it’s C797S. That’s the site where osimertinib covalently binds the EGF receptors, so that makes sense. Researchers found that mutation in about 7% of patients in the FLAURA trial, and there was a higher proportion in the second-line study. It’s important to find out whether it is cis or trans, because it can dictate what an optimal next-line strategy would be, with really good clinical data. A lot of it is based on in-vitro data and case reports for small case series at the moment.

Other mechanisms of resistance to up-front osimertinib include MET amplification, or amplification of other oncogenic drivers like KRAS, HER2 [human epidermal growth factor receptor 2], BRAF, PIK3CA—activating mutations in some of the usual suspects: MET, HER2, FGFR, BRAF, KRAS, and PIK3CA. There’s also amplification for a subset of patients with cell cycle regulating genes, for example CDK6 or other cyclin Ds. For a small portion of patients, there are oncogenic-driving fusion mechanisms including ALK, RET, and BRAF, and that can be the case for up to 10% of patients. Importantly, though, patients can transform from having non–small cell lung cancer to small cell lung cancer or squamous cell lung cancer. That really underlines the importance of doing biopsies for these patients and not wholly relying on ctDNA [circulating tumor DNA]. A biopsy at progression is crucial for these patients, because if someone has a small cell transformation, we’re going to treat them very differently from if they didn’t have that. There are lots of complexities, and essentially what we’re doing is excluding a relatively uncommon type of lung cancer—it certainly is in the United Kingdom—into many more diseases, which makes clinical trials challenging in this space, but there are plenty of clinical trial pitches underway to try to find out the optimal way of chasing down a precision medicine paradigm for these patients.

There are dozens of these trials, but I would just like to highlight the ORCHARD study, which is a phase 2 study that matches patients to—usually—a combination-based regimen: osimertinib plus something else, depending on the mechanism of toxicity. For example, if the patient has a MET amplification, they should get osimertinib plus savolitinib; if they have an EGFR gene amplification, they get necitumumab. If they have an EGFR C797S mutation, they get additional gefitinib because, somewhat unexpectedly—well, if you look at the structure then it works, but you can resensitize patients to these first- and second-generation TKIs [tyrosine kinase inhibitors] based on which allele the C797S mutation arises. There are many strategies, and it will be challenging to do phase 3 trials within this space in terms of recruiting the number of patients required. Our focus must be on large phase 2 studies with rationally matched therapies, but of course there are biomarker-agnostic approaches as well. That’s when you’re forgetting about finding the genomic biomarker, and you can think about other strategies to treat these patients. In the IMpower150 trial, which has already been mentioned, researchers utilized chemotherapy with a platinum: paclitaxel plus bevacizumab in addition to atezolizumab. That’s a regimen that has shown some efficacy in the treatment of EGFR-mutant disease, so that gives us a bit of promise that precision medicine may not be everything within this disease’s treatment. We know that chemotherapy itself has a role, so there are promising avenues for these patients, but testing is key, and having clinical trials available for these patients is important not only for them but also for determining the best evidence to treat future patients.

Neal Navani, MRCP, MSc, PhD: Thanks, Gary. Solange, you’ve just published a fantastic review in this area in Nature Cancer that I recommend everyone read. It’s a great piece of work. Tell us, based on that, do you think everybody who’s progressing on osimertinib should have a biopsy? Is that something that we’re starting to need to build into our clinical practice routinely, or do you think liquid-based tests are enough in the treatment of stage IV disease?

Solange Peters, MD, PhD: It strongly depends on what’s available down the road for you. Meaning, do you have a trial similar or smaller than the ORCHARD trial, which will allow you to get access to some MET inhibitor or some other potential targeted therapy in case you identify such a mechanism? That’s really the critical question. If you don’t have any of this accessibility to trials or even to out-of-indication, off-label molecules, there is this transformation into small cell. In all the literature we’ve been seeing, apart from 2 series from the United States, it’s important to say that liquid biopsies were used, including in the FLAURA trial, for example. It meant that we probably underestimated the transformation into small cell disease, so you might not redo a biopsy and go ahead with chemotherapy, or with a chemotherapy that will be more adaptive to a genome compared with a small cell component. That’s the negotiation—the mitigation—you accept if you don’t biopsy.

I’m not sure if we should change this so much, because what I’ve been seeing with my small number of patients with small cell transformation is, in this specific small cell—this population consists of never-smokers—is an acquired TP53 and RB1; nothing works. You can give platinum-etoposide. Well, it’s not the small cell that has this 80% response rate. Small cell is bad, but it’s even worse than the usual small cell. To which extent it really changes to give platinum, pemetrexed, pembrolizumab, or nivolumab 150 mg vs a platinum-etoposide, I’m not even sure. To answer your question, I’d say that there are some mandatory biopsies. This 1 is more for the purpose of research, for understanding, but it might be that—at the time it was being used routinely across the globe—it changes so much in terms of the patients’ perspective. Perhaps it’s not mandatory; it’s for curiosity.

Neal Navani, MRCP, MSc, PhD: Thanks, Solange. Christian, what’s your view? You talked about doing biopsies on patients who had progression after first-line osimertinib. Is that part of your routine practice, as well as liquid biopsies, to look for phenotypic transformation?

Christian Grohé, MD: If we can obtain tissue from the site of recurrence of the progressive tumor, we would go for the tumor specimen taken from the original site, or the site of recurrence. If that’s not feasible, or technically not feasible, we go for liquid biopsies. Recognizing that, we’re still trying to understand what’s going to happen, and it gives us some information about that. What Gary pointed out is that we’ll see transformation into other disease entities. We see certain patterns of escape mechanisms, but what’s important for us to understand is that in that context, the liquid biopsy is quite helpful. Do we see—in the liquid biopsy, with a certain penetrance in terms of allele frequency—the original clone, or does the clone disappear? Did it disappear? We very often have the first cases of patients who have been operated on early on, a couple of years ago, who had an EGFR mutation and now have recurrence. It’s not an easily accessible recurrence. Through liquid biopsy, we can see that, for these patients, the clone has disappeared. Obviously, if they have a tumor, they have a different tumor, and for that reason we have to access that instead of using or reaching these patients with a TKI. In terms of developing diagnostic algorithms in driver-mutated lung cancer after failure, it’s very important to standardize the approach to understand the unmet need and what the pitfalls look like.

Neal Navani, MRCP, MSc, PhD: Do you have any concerns about the sensitivity of liquid biopsy in this scenario?

Christian Grohé, MD: In terms of quality, it’s a very good question. We had to do platinum for a very long time, with the same team, who has developed a very fine approach. So far, we’ve discussed that with the molecular pathology board every week in detail to understand what’s going on. Often, we compare the results from the liquid biopsy with the results we have from the tumor specimen originally, and there can be results that are different, but still at the end we conclude the best therapeutic approach we would like to recommend. Still, this is a learning curve—you’re right, and in terms of pushing that field, that’s going to be helpful for all of us.

Transcript Edited for Clarity