Therapeutic Options Developing for Non-Driver NSCLC, But Challenges Remain

Corey Langer, MD, discusses KEYNOTE-189, the current arena in NSCLC treatment, and the obstacles that remain in effectively treating patients with non-driver mutations.

Corey Langer, MD

While challenges remain for patients with non-driver non-small lung cancer (NSCLC), there continues to be ongoing trials that are evaluating and testing various immunotherapy and chemotherapy combinations in the frontline setting, according to Corey Langer, MD.

In the ongoing phase III KEYNOTE-189 trial, the efficacy of the PD-1 inhibitor pembrolizumab (Keytruda) combined with platinum/pemetrexed chemotherapy versus chemotherapy alone is being studied in the frontline setting for patients with nonsquamous NSCLC (NCT02578680).

This is a follow-up study to cohort G of the KEYNOTE-021 trial, in which the pembrolizumab triplet elicited an objective response rate of 55% compared with 29% with the chemotherapy agents alone (P = .0032). The FDA approved the regimen in May 2017 based on these findings.

Moreover, practicing clinicians have the standard therapeutic option of the angiogenesis inhibitor bevacizumab (Avastin) in this patient population. In 2006, the FDA granted approval for a labeling extension for bevacizumab administered in combination with carboplatin/paclitaxel as frontline treatment of patients with unresectable, locally advanced, recurrent, or metastatic, nonsquamous disease.

OncLive: What is the prevalence of patients with non-driver NSCLC?

What are the current treatment strategies for nonsquamous NSCLC? Also, what are the considering factors for bevacizumab?

In an interview with OncLive, Langer, who is director of thoracic oncology at the University of Pennsylvania, provided further detail of the goals and implications of this clinical trial. Moreover, he discussed the current arena in NSCLC treatment and the obstacles that remain in effectively treating patients with non-driver mutations.Langer: The incidence of patients with non-driver mutations in NSCLC, specifically adenocarcinoma, is close to 70%. In squamous cell carcinoma, it’s 95% or higher. The vast majority of individuals we treat with advanced NSCLC, either metastatic at diagnosis or recurrent, do not have driver mutations or translocations. For those individuals, specifically with nonsquamous disease, the standard options generally include pemetrexed with carboplatin or a taxane with carboplatin. Now, [we have] the option of pembrolizumab combined with a platinum-based combination and, of course, bevacizumab combined with a platinum-based combination. The standard approach for nonsquamous NSCLC, at least in 2017, involves a platinum combination in individuals who do not have an oncogenic driver. Usually that platinum combination in the United States is carboplatin and pemetrexed. In individuals who are not candidates for pemetrexed, it is carboplatin plus a taxane either every 3 weeks or weekly.

Up until recently, in individuals who not have a history of hemoptysis or untreated brain metastases, I would generally add bevacizumab onto the platinum combination of either carboplatin and paclitaxel or carboplatin and pemetrexed. With the recent approval of pembrolizumab in this setting—specifically pembrolizumab in combination with pemetrexed and carboplatin—our therapeutic options have expanded. However, they have also gotten a bit more confusing.

Pembrolizumab was approved based on a randomized phase II trial that compared the combination of pembrolizumab with chemotherapy to just chemotherapy alone. Bevacizumab was not included in the control arm. We have no direct head-to-head comparisons of pemetrexed/carboplatin plus bevacizumab versus pemetrexed-carboplatin plus pembrolizumab.

That is really the root of one of my dilemmas. There are some patients who have an autoimmune disease or other contraindications to pembrolizumab where I will certainly and preferentially use bevacizumab.

The other interesting population are those with oncogenic drivers—be it an EGFR mutation or ALK translocation. Here, immunotherapy does not work that well. If you look at the randomized phase III trials in the second-line setting that compared single agents with docetaxel, the 1 group that consistently failed to derive any benefit were those with oncogenic drivers. The response rates were quite low: they were 3% or 4%. Survival advantages were not seen in that group. In fact, in some of these studies, there was a survival advantage for chemotherapy. Inevitably, those individuals will develop disease progression on the original tyrosine kinase inhibitor (TKI).

What are some ongoing clinical trials with therapies for non-driver NSCLC?

At that point, chemotherapy is the standard. In my mind, chemotherapy, specifically a carboplatin combination with bevacizumab is the standard for that population. I would much sooner use bevacizumab in those with oncogenic drivers once their disease has progressed on the initial TKI ahead of using an immunotherapy, such as pembrolizumab, nivolumab (Opdivo), or atezolizumab (Tecentriq). Whereas, in a wild-type patient…I would preferentially use immunotherapy in that setting.There are trials specific to squamous and nonsquamous patients. Then, there are the immunotherapy combinations versus chemotherapy. Since 2015 and 2016, there have been a number of prospective randomized phase II, and now phase III, trials that are looking at immunotherapy “cocktails,” if you will, or combinations in the frontline setting. Many of these trials—the majority, in fact—continue to use the chemotherapy backbone. It is standard chemotherapy [with or without] the new immunotherapy drug—be it nivolumab, pembrolizumab, or atezolizumab—both in nonsquamous and squamous cell NSCLC.

One of the more interesting trials is KEYNOTE-189, which is essentially a phase III follow-up of cohort G of the randomized phase II KEYNOTE-021 trial. Looking at pemetrexed and carboplatin alone or combined with pembrolizumab, it has a primary endpoint of disease-free survival. It has crossover built into the trial; hence, overall survival is not a legitimate primary endpoint. It is still an important secondary endpoint. This trial will either confirm or refute the benefits that we have seen in cohort G, which led to the approval of that 3-drug combination.

What are some ongoing challenges with the non-driver population?

There is a similar trial looking at a taxane/platinum combination in squamous cell NSCLC. Atezolizumab, which is a PD-L1 inhibitor, is being looked at prospectively in phase III trials, again, with several different chemotherapy backbones. There are a number of trials now that are looking at immunotherapy combinations, particularly CTLA-4 inhibitors and either PD-1 or PD-L1 inhibitors together versus chemotherapy. Those trials are truly exciting. I have some degree of skepticism as to whether immunotherapy combinations will actually prove superior to chemotherapy in this setting, but I would not be the least bit surprised if they prove equivalent.There are a number of major challenges that we are dealing with in the non-driver population. First off, the absence of an oncogenic driver limits the therapeutic options. Unfortunately, historically, median survival in this group is inferior to those with oncogenic drivers.

There is a subpopulation, maybe 5% or 10%, where immunotherapy not only doesn't work, it may work to the patient's detriment. We have seen this phenomenon of accelerated or hyperprogression in some individuals, which is quite frightening. An individual might have had progressive disease, they are treated with immunotherapy, and then their disease takes off. We have no way of identifying those patients.

Intriguingly, individuals who have been exposed to immunotherapy—some of whom include those with hyperprogression—and are rechallenged with chemotherapy may actually do fairly well. We may actually be seeing some abscopal or remote effect that makes chemotherapy work better.

Subsequently, we need to figure out what the optimal sequence is. We need to do a far better job of identifying individuals who are more likely to respond to immunotherapy agents. PD-L1 testing has become our new standard in the upfront setting. It identifies the 30% or so with high levels of expression who do better with immunotherapy—specifically, pembrolizumab—over chemotherapy alone. I am convinced that there are other markers that can predict who is likely to benefit; tumor mutation burden is just 1 example. Now, with next-generation sequencing, we can identify the number of mutations that each individual has. There are retrospective studies that show that the mutation load can correlate with the response benefit to these new agents.

For those who are lucky to respond, we have no idea how long to continue these treatments. There are increasing data that show that treatment beyond 2 years probably doesn’t necessarily benefit these individuals any further. There are now trials that are looking at those who have had a response that are comparing 1 year to 2 years with indefinite treatment— obviously, with the option of resuming treatment should disease ultimately progress.

These are all unanswered questions. For those who have enjoyed a response, that's a good dilemma to have. The bigger challenge is with those who don't respond all that well. Some sort of combination immunotherapy strategy or unique combinations between immunotherapy and chemotherapy could overwhelm or overcome intrinsic resistance. That, to me, is probably the biggest challenge.

Langer CJ, Gadgeel SM, Borghaei H, et al. Carboplatin and pemetrexed with or without pembrolizumab for advanced, non-squamous non-small-cell lung cancer: a randomised, phase 2 cohort of the open-label KEYNOTE-021 study. Lancet Oncol. 2016;17(11):1497-1508.