Novel Agents Continue to Drive Treatment of NSCLC With EGFR, MET, and ALK Abnormalities

Targeted agents for non–small cell lung cancer (NSCLC) harboring EGFR, MET, ALK, and other mutations continue to emerge in the treatment paradigm, adding importance to the use of genomic testing, including in patients with early-stage disease, according to Jun Zhang, MD, PhD.

“We know that if we combine [liquid and tissue biopsies] together, that we can increase the odds by about 20% of finding targetable mutations. If either [biopsy] gives us [an indication of a patient] having those targetable mutations, then we feel comfortable to initiate targeted therapy,” Zhang said in an interview with OncLive^® following an Institutional Perspectives in Cancer (IPC) webinar on lung cancer, which he chaired.

In the interview, Zhang discussed key clinical trials utilizing immunotherapy in NSCLC, the continued importance of biomarker testing to help inform treatment decisions, and targeted therapies for EGFR, MET, and ALK mutations. Zhang is an associate professor in medical oncology at the University of Kansas and a medical oncologist, Division of Medical Oncology, Department of Internal Medicine, Department of Cancer Biology, University of Kansas Medical Center.

OncLive^®: In your presentation, you discussed the use of durvalumab (Imfinzi) in the phase 3 PACIFIC trial (NCT02125461). Could you expand on what was learned from these data?

Zhang: Durvalumab maintenance prolongs progression-free survival [PFS] and overall survival [OS]. My personal thought is durvalumab is curing more cancer. There is also a post-hoc analysis of some of the subgroups [showing that] patients with EGFR or ALK mutations don’t seem to benefit much from durvalumab, which is consistent with what we see in the metastatic setting.

For PD-L1 expression, we do have to interpret data cautiously. In the post-hoc analysis, patients with PD-L1–disease expression did not seem to derive significant benefit compared with the control arm. However, I want to make it clear that this study was not designed in this way. In fact, many of the patients did not have initial PD-L1 testing when they started treatment.

We also know that with chemotherapy or radiation therapy, or especially that combination, there’s a good chance that we can upregulate PD-L1 expression of tumors. Because of that, I don’t think we should withhold treatment, especially durvalumab, for those patients with negative PD-L1 expression.

Patients with unresectable disease, historically, need more improvement. The PACIFIC study is the first study using an immune checkpoint inhibitor that demonstrated survival benefit in the curative-intent setting. This is very important for any patients who are not candidates for surgery. I would recommend chemotherapy/radiation therapy before immunotherapy and maintenance. We can learn a lot from that study, especially questions [about] whether those patients who, for example, don’t have a good performance status, are ideal candidates for chemotherapy/radiation therapy. Is there any value if we give durvalumab in that setting?

There are ongoing clinical trials, but in the real world, sometimes we [study] sequential [treatment]. There was 1 [recent] study from China called GEMSTONE-301 [NCT03728556] that allowed sequential chemotherapy/radiation therapy followed by the immune checkpoint inhibitor [sugemalimab]. We can probably [conclude] from the data from [GEMSTONE-301] that for those patients, even if they’re not ideal candidates for chemotherapy/radiation therapy, if we [treat sequentially], there might be some value to give immune checkpoint inhibitors after [chemotherapy/radiation therapy].

Could you expand on what was learned from the phase 3 CheckMate 227 (NCT02477826) and CheckMate 9LA (NCT03215706) trials?

We also discussed 2 studies in the metastatic setting, which were CheckMate-227 and CheckMate 9LA. My colleagues and I are comfortable using pembrolizumab [Keytruda] and cemiplimab-rwlc [Libtayo] in these patients with PD-L1 expression of more than 50%, and when it’s lower than that or [PD-L1 is] negative, then [clinicians] are comfortable using a combination of chemotherapy with immunotherapy [such as] pembrolizumab.

The question is always [about] where we should place these 2 regimens. For example, the phase 3 KEYNOTE-024 trial [NCT02142738] [established the use of] single-agent pembrolizumab for patients with a high PD-L1 expression. We have the phase 3 KEYNOTE-189 trial [(NCT02578680) with the combination of pembrolizumab and chemotherapy], and then we have the phase 3 KEYNOTE-407 trial [NCT02775435] for squamous NSCLC, again with the combination [of pembrolizumab] and chemotherapy.

There’s no head-to-head [trial], but the beauty is that we have 4-year data from CheckMate-227. We have 5-year data from KEYNOTE-024. We have 2- or 3-year data for KEYNOTE-189 and KEYNOTE-407. These [data] give us a glimpse about [each regimen’s long-term] efficacy in the [metastatic] setting.

What we found is for patients with high PD-L1 expression of more than 50%, if we look at 2-, 3-, or 4-year OS, [the rates] seem to be comparable with [those from] CheckMate-227 when you use dual agents. It makes sense [that clinicians] prefer single-agent pembrolizumab vs dual-agent ipilimumab [Yervoy] plus nivolumab [Opdivo]. We know that [when] we have 2 immunotherapy [agents], there is a higher chance of developing [immune-related] toxicities. If the survival is similar, it makes perfect sense that we give single-agent [pembrolizumab]. We probably need more time to tell [whether single-agent or dual-agent treatment has] more durable responses.

When we look at patients who are PD-L1 negative or [have] less than 1% [expression], we saw especially for patients with squamous NSCLC when we compare 2-year [data] between the KEYNOTE-407 vs CheckMate-227, that CheckMate-227 seems to show a better outcome than KEYNOTE-407. [These data] suggest that for patients with squamous histology and PD-L1 expression less than 1%, there’s good value for [nivolumab plus ipilimumab].

For adenocarcinoma, [these data are laregly] comparable. I see maybe numerically better [outcomes] in some of the 2- or 3- year data [from KEYNOTE-189 and KEYNOTE-407], but they are largely comparable. We will [continue to see updated] data from the CheckMate studies. It makes perfect sense that we gave single-agent pembrolizumab or cemiplimab for patients with greater than 50% expression of PD-L1, but for patients with less than 1% PD-L1 expression, especially those with squamous histology, I favor the CheckMate-227 regimen of ipilimumab plus nivolumab. For patients [with PD-L1 expression] between 1% and 50%, if the patient has no interest in chemotherapy, then [ipilimumab plus nivolumab] is a reasonable alternative approach.

The beauty of [CheckMate-227] is this is probably the first study that has the longest follow-up with dual-agent immune checkpoint inhibitors targeting the PD-L1 pathway and CTLA-4. They are different nodes in the immune cycle. If we agree with the design of CheckMate-227, [the] CheckMate 9LA [regimen] makes perfect sense, especially for those patients with bulky tumors [where] you want to give 2 additional cycles of chemotherapy to achieve cytoreduction [and relieve patient symptoms]. Whether this will translate to survival benefit, no one knows for now. We only have 2-year data for CheckMate 9LA.

How do driver mutations affect treatment decisions for patients? How do you approach testing for these alterations?

We discussed whether we should do both liquid biopsy and tissue biopsy, and everybody agreed that we should do liquid biopsy [in addition to tissue biopsy] for a couple reasons. Number 1 is the fast turnaround time. Number 2 is in case tumor tissue is not sufficient, then at least we have another way to know the status of mutations. We also know that driver mutations detected from liquid or blood are also functional. In other words, if you give targeted therapy, patients respond the same way as if they had the mutation in the tumor tissue.

In my practice and that of my colleagues, especially for those patients with NSCLC, we prefer to do both at the same time.

Our second question is about the whether we should take everybody for next-generation sequencing [NGS]. We did discuss that it is more cost effective to use [polymerase chain reaction (PCR) to test for] 1 single gene and 1 mutation. But now we have 9 targets with mutations that have FDA-approved targeted therapies, especially, for example, EGFR. I’ll use EGFR exon 20 insertion [mutations] as an example. Theoretically, we can have more than 100 variants that can insert at any place. Any commercial PCR-based approach would probably not be ideal, and even if we want to do PCR, like designing primers to flank the exon on 2 sides, we still have to sequence a whole PCR product. In that case, if insurance companies can cover it, I certainly prefer NGS, and this will also give us other genetic mutations that might open the opportunity for clinical trials for our patients.

How do you anticipate genomic testing will continue to evolve in lung cancer?

We discussed what other modalities we may [see] in the future. In the old days we used fluorescence in situ hybridization [FISH] analysis for fusions, but now we can use RNA sequencing. RNA-based NGS is good to find those insertions or truncated deletions, as well as fusion products. In the future, immunohistochemistry [IHC] will be an essential component. We already saw that, for example, with PD-L1 expression. In the future, we will have new drugs—for example, antibody-drug conjugates that target cell surface proteins.

As an example, Prakash Neupane, MD, of the University of Kansas Medical Center, presented the [phase 1 CHRYSALIS trial (NCT02609776)] that combined amivantamab-vmjw [Rybrevant] and lazertinib [Leclaza] in patients with EGFR exon 20 insertion [mutations]. Retrospectively, those patients by IHC, if they had good expression of EGFR or c-MET, their response was much higher. In fact, the chance [of response] was about 90%, confirming the value of IHC.

We also [have] potential biomarkers from the [patient] level, not only necessarily [from the] tumor. We know that, ultimately, if we want to [improve] survival, not only do we have to get rid of the tumor, but, also, we have to get support from the [patient]. Potential [biomarkers] are, for example, proteomics from blood. The immune profiling is based on peripheral blood mononuclear cells in blood, even microbiotics, as well. Those are all the potential future biomarkers that we can study.

Finally, we discussed the value of biomarker testing in early-stage [disease] for everybody, not just in the metastatic setting. For example, the phase 3 ADAURA trial [NCT02511106] [showed] us that in patients with EGFR mutations, if you give an EGFR TKI you can prolong PFS. The phase 3 IMpower010 trial [NCT02486718] showed that if [patients] have positive PD-L1 expression, then they may benefit from atezolizumab [Tecentriq] as immunotherapy in the adjuvant setting.

In the future, we are going to have more studies for [checkpoint inhibitors] in the neoadjuvant or adjuvant setting, and biomarkers in the early stage will be valuable. Now the question is whether we should just go with PCR or NGS, and everybody agreed that from a biology or science perspective, NGS certainly is better because [it provides] more information. But from the real world, if we cannot provide additional treatments, for example, at this moment, only the EGFR TKI osimertinib [(Tagrisso) is approved] for patients with EGFR [mutations], but not other mutations, and if the insurance is not paying for [NGS], then we are okay for just targeted sequencing based on PCR.

Expanding on the presentation from Prakash Neupane, MD, on EGFR- and EGFR exon 20–targeted therapies, we have osimertinib and now 2 specific exon 20 agents in mobocertinib (Exkivity) and amivantamab. What is important to know about these agents and their applications in practice?

In the old days, we only knew those common mutations, such as exon 19 and exon 21. Occasionally, some people were aware of exon 18, what we call EGFR sensitizing mutations, but now we know that there’s also targeted EGFR mutations, which we call exon 20 insertions. Patients with exon 20 insertions tend to have poor prognosis. At the same time, they do not respond to those EGFR TKIs from different generations.

The CHRYSALIS study is valuable in the sense that it demonstrates the value of amivantamab in this setting. Numerically, it seems like amivantamab has higher response rates compared with mobocertinib, but mobocertinib has a much longer duration of response. These 2 drugs take hold of the same EGFR exon 20 insertion [mutation] but from different mechanisms. Amivantamab is a monoclonal antibody targeting both EGFR and MET, as we know that MET can be the dimerization polymer for EGFR. All those members in the EGFR family have to have a dimerization, but they activate down from a single pathway. The c-MET [pathway] could be 1 of the dimerization polymers, and we also know that the activity of c-MET could be another mechanism of resistance. Amivantamab is a monoclonal antibody targeting EGFR, whereas mobocertinib is a TKI. They [attack] from 2 different angles, and I’m glad to see that we have 2 agents to target this EGFR mutation.

In the future, there will be more studies combining, for example, the EGFR TKI lazertinib with amivantamab. There are studies ongoing, and mobocertinib is also [involved] in trials with different novel combinations, and hopefully we will see a better response in this patient population.

Dr Neupane also discussed the well-known phase 3 FLAURA study [NCT02296125] using osimertinib, a third-generation EGFR TKI in the first-line setting. Again, these are positive data, and [the results] confirmed that there’s great therapeutic value if we start using third-generation TKIs in the first-line setting. In the community, some people still prefer to rotate from first- to second- and then to third-generation [TKIs]. Part of the reason [for that strategy] is if patients develop resistance to the first-generation [TKIs], then we know that there is a 50% chance that they have a T790M mutation, and then we know they are going to respond to osimertinib.

That is a strong argument, but at the same time, [FLAURA] tells us that about 30% of patients never had a chance to cross over, once they failed on their first-line EGFR TKI, to receive third-generation osimertinib.

Therefore, in my practice, I favor giving the best [treatment options] to my patients [in the first line]. It doesn’t make sense that they have to suffer. Another thing is, we have research ongoing every day. We’re now having more options. Dr Neupane also discussed the phase 3 MARIPOSA study [NCT04487080] for those patients with EGFR mutations in the first-line setting, [using] the combination of lazertinib, which is a third-generation EGFR TKI, along with amivantamab, and we are going to compare that approach directly with osimertinib to see [which treatment elicits] better responses. There are some other studies using amivantamab in the second-line setting for [patients who] fail an EGFR TKI. There is also a cohort from the CHRYSALIS study [examining lazertinib plus amivantamab].

I would not be worried about dealing with unknown consequences when patients develop resistance to osimertinib. We have good data suggesting that we know the resistance mechanisms [in that setting]. [For example], c-MET amplification could be 1 of the resistance mechanisms, [plus] C797S mutations, and we do have a few studies with promising data that [can help us] tackle these problems when patients develop resistance. For those reasons, I do not [hesitate] to use osimertinib in the first-line setting.

Now that more patients with EGFR mutations are receiving osimertinib in the adjuvant setting, if ultimately those patients develop metastatic disease, how will that affect osimertinib’s use in the frontline metastatic setting?

We know that cancer cells have what we call plasticity. They always find a way to escape, especially when you give targeted therapy just focusing on 1 pathway. It is not surprising that cancer will find and utilize another pathway to escape or [that patients] develop a secondary mutation, or, commonly, they develop EMT, which is what we call epithelial mesenchymal transition. That is all a possibility. The resistance mechanisms [we’ll see in the adjuvant setting] will [not] be different compared with [those seen in patients in the metastatic setting]. The biology is very similar, if not exactly same, so there will be a huge overlap of resistance mechanisms.

Based on that, what we are going to [use to] tackle patients with metastatic cancer and resistance to osimertinib can also be utilized for that patient population. Having said this, in my practice, I will always like to have another biopsy. It could be both [liquid and tissue biopsies]. It could be at least liquid biopsy to understand the clonal evolution. Sometimes we see the patients that completely lose the mutation, and in those patients, maybe we should consider chemotherapy. Some patients did develop through some signal amplification. In their case, we may want to add a c-MET inhibitor, and in that scenario, maybe adding amivantamab could be an interesting thought.

Some other patients may switch to [having] some other mutations. [For example], 1 of our patients switched to a BRAF mutation. It is odd, but then the patient responded to BRAF-targeted therapy. There are also occasions [where] we saw some patients have what we call tissue transdifferentiation from non–small cell adenocarcinoma to small cell [adenocarcinoma], setting a retinoblastoma wall, which is tumor suppressor loss. In that case, we’ll probably have to consider a similar chemotherapy regimen that we use for small cell lung cancer.

In the presentation by Chao H. Huang, MD, FACP, of the University of Kansas Medical Center, on targeted therapy for NSCLC, what types of mutations are prevalent in this space?

Dr Huang provided a good summary of two studies with RET inhibitors. In lung cancer, we’re talking about the RET fusion protein. We talk about that mutation [for thyroid cancer], but [it is more common] for NSCLC. [Dr Huang] discussed the phase 1/2 LIBRETTO-001 study [NCT03157128] and the phase 1/2 ARROW study [(NCT03037385) in RET-mutated NSCLC]. LIBRETTO-001 is using selpercatinib [Retevmo], whereas ARROW is using pralsetinib [Gavreto]. Both [trials] demonstrated an excellent response rate and excellent intracranial activity.

Clinically, we use both [selpercatinib and pralsetinib]. During the meeting, I asked Dr Huang: How are you going to sequence [these agents]? He explained that he had 1 patient who started with selpercatinib but developed some fever and other [adverse] effects [AEs] and was switched to pralsetinib. When you look at numbers [between selpercatinib and pralsetinib], maybe there is a little difference [in response rates]. One has a high response rate with a lower complete response [rate], and the other one is the opposite, but both have superior intracranial activity.

For me, there is not much difference [in efficacy, so] I look at the toxicity profile. For example, 1 agent has a higher chance of pneumonitis, and the other has a higher chance of QT prolongation. This is something that we have to put into consideration.

Dr Huang also gave a comprehensive review of ALK inhibitors. He discussed 3 second-generation TKIs in comparison to either the first-generation crizotinib [Xalkori] or chemotherapy, namely ceritinib [Zykadia], brigatinib [Alunbrig], and alectinib [Alecensa]. He discussed using the third-generation [ALK inhibitor] lorlatinib [Lorbrena]. We discussed how you sequence [lorlatinib], and this happens all the time. In general, we are comfortable using any of the second-generation [ALK] inhibitors. Though, I prefer alectinib. Some other [clinicians] probably prefer brigatinib. It depends on personal experience, and the most important thing is the toxicity profile [for patients].

Regarding lorlatinib, the phase 3 CROWN trial [NCT03052608] demonstrated a superior hazard ratio [vs crizotinib]. If we look at all the previous studies, the CROWN study probably gave us the lowest hazard ratio compared with chemotherapy. If we want to justify [lorlatinib] as a first-line treatment, we have good reason.

Having said this, again, in general, patients respond nicely to second-generation [ALK inhibitors]. They are well tolerated, especially alectinib. There are some issues with crizotinib with gastrointestinal toxicity. Brigatinib occasionally will have pneumonitis, but alectinib is well tolerated, and the duration of response is generally long. Certainly it's a good choice, but at the same time, I will not hesitate to give lorlatinib in the first-line setting.

For example, one of my patients, a young woman with kids, has metastases everywhere with brain metastases, as well. I gave the patient lorlatinib, it was tolerated well, and the tumor melted away like magic. I have to do is watch her [lipid] profile though, because [lorlatinib can] increase low-density lipoprotein and triglyceride. [We] have to be careful about that. Otherwise, first-line lorlatinib [is a good option]. Dr Huang briefly touched on fourth-generation [ALK inhibitors], which are ongoing in clinical trials, targeting gatekeeper mutations. We hope that we are going to see more exciting data [in fourth-generation ALK inhibitors] in the next few months or years.

Is there any ongoing or planned research at the University of Kansas in any of these areas that you would like to highlight?

We are focused on investigator-initiated trials. Patients with PD-L1 expression less than 1% do not derive benefit from immune checkpoint inhibitors very well. Because of that, we are doing an investigator-initiated clinical trial that has not opened yet, where we are trying to combine the multikinase inhibitor cabozantinib [Cabometyx] with atezolizumab. The whole idea behind it is [that] cabozantinib is a multikinase TKI [that hits] VEGF, and we know that VEGF activation counters immune suppressive phenotypes for tumor microenvironments. We hope that we can turn this immune cold to immune hot tumor and sensitize it to immunotherapy.

I’m also actively studying the microbiome to predict response to immunotherapy, and, more importantly, also to predict the immune-mediated toxicities to immunotherapy. We don’t want to see patients that are not developing therapeutic benefit and, at the same, suffer from unnecessary immune-mediated toxicities. [One such study] is funded by AstraZeneca, where we are looking in the specific population [receiving] durvalumab once they finish chemoradiation therapy. We are harvesting the microbiome from different organ systems to see whether we can predict the immune-mediated toxicity from durvalumab. If that’s the case, then we can better identify those patients and avoid unnecessary AEs, especially pneumonitis, which sometimes can be life threatening.

In the targeted therapy area, we planned a phase 1 trial using a novel compound targeting both regular EGFR [mutations], as well as EGFR exon 20 insertion and HER2 exon 20 insertions. [There are] common features between EGFR exon 20 and HER2 exon 20 insertions, and we hope that this new compound will have some interesting findings.

Is there any ongoing or planned research at the University of Kansas in any of these areas that you would like to highlight?

The peer-to-peer discussion is good. We are limited in a sense [in] our knowledge, capacity, or ability to interpret data, as well as our personal experience when we treat our patients. The exchange of ideas or experiences is always valuable, and I hope that we can have regular activity like this, to go through the important trial data, exchange our ideas, give feedback to [drug] manufacturers, and also spread the new good news to other health caregivers as well as our patients.

One other thing that is important is, I always tell my patients, especially for those patients with cancer, what we call the standard of care of treatment today for you, primarily or largely is based on clinical trials. Consider clinical trials as a valuable resource. Not only do you have access to cutting-edge science and technology, but at the same time, you have the chance to contribute to science and benefit future patients with the same disease. [At] the end of the day, we have to realize that we are all connected to each other. We are benefiting from other people and hopefully we can benefit others.