Immunotherapy Revolutionizes Squamous NSCLC, But Biomarker Enhancement Needed

Chad Pecot, MD

Chad Pecot, MD

The introduction of immunotherapy to the treatment landscape for stage IV squamous non—small cell lung cancer (NSCLC) has been revolutionary, with thousands of ongoing international trials exploring new combinations, said Chad Pecot, MD. However, better models for evaluating these regimens are desperately needed, as are effective biomarkers for immune response.

“Immunotherapy has really changed the game in lung squamous carcinoma, said Pecot, who is an assistant professor, Department of Medicine, at the University of North Carolina (UNC) Lineberger Comprehensive Cancer Center. “There is such a fervor right now to try and find the best biomarkers [for response]. All over the world, many, many groups are trying to determine what the best biomarker is for immunotherapy, which kind of implies how bad the biomarkers are currently.”

One biomarker that has been utilized in the space is PD-L1 expression, and although it can help predict some patients who might benefit from immunotherapy, lack of its presence does not necessarily indicate that a patient will not derive benefit from checkpoint inhibition.

“Anecdotally, myself and many others have had many patients receive immunotherapies who have a zero PD-L1 [expression] and still have a great response [to this approach],” said Pecot. “You don't want to avoid giving a patient a treatment that amazing [if they may benefit].”

In an interview during the 2019 OncLive^® State of the Science Summit™ on Non—Small Cell Lung Cancer, Pecot discussed practice-changing immunotherapy research that has read out in the space, emerging biomarkers under investigation, and logical combinations that beg for further exploration.

OncLive^®: How has the introduction of immunotherapy impacted stage IV squamous NSCLC treatment?

Pecot: In North Carolina, I take care of patient populations [who are] very heavy smokers. We see a huge population of patients with squamous NSCLC. I came to UNC 6 years ago, and when I first started, we only had a few different chemotherapy options available, and nothing very exciting beyond that. There were no real targeted therapies in store for squamous NSCLC. In fact, there were several trials at the time looking at targeted therapies in squamous disease, and they didn't pan out very well.

Then, immunotherapies came along and completely changed everything; we now have quite a few patients who are long-term survivors of metastatic lung squamous carcinoma, [surviving] sometimes for years at a time. Therefore, yes, it has been pretty revolutionary.

What key trials did you highlight in your presentation at the State of the Science Summit™?

The first trial that I presented was KEYNOTE-407; it's a very exciting trial. There was clear improvement in OS [observed], which is kind of the real litmus test [when it comes to evaluating regimens]: does OS improve? It clearly did when you added pembrolizumab (Keytruda) to standard chemotherapy that we use in squamous lung carcinoma—in this case, carboplatin and paclitaxel or nab-paclitaxel (Abraxane). This trial was interesting in that it actually allowed crossover, so when patients who didn't initially receive the immunotherapy started experiencing tumor growth, they were able to start receiving immunotherapy.

To me, there are a couple of things that are really interesting in this trial. First, it didn't matter how much PD-L1 expression was in their tumor; all patients benefitted. Even patients who had no PD-L1 expression in their tumor had a very nice OS benefit by adding immunotherapy. The other thing that's actually interesting is, although they allowed crossover, there was still an OS improvement if the immunotherapy was received earlier, meaning that the patients who were enrolled to the immunotherapy arm from the get-go had the best survival.

The patients who didn't initially receive immunotherapy and then eventually received it still benefitted from the drug; their survival was just not as [prolonged]. That implies that receiving immunotherapy early matters, independent of how much PD-L1 expression a tumor has. That is probably the real take-home message from KEYNOTE-407. [Furthermore], patients handled the medicine well; adding immunotherapy didn't add much toxicity and patients lived longer, independent of what their PD-L1 expression showed. Therefore, that is the new standard of care for squamous cell lung carcinoma.

The [IMpower131] study is still early; the data are not as mature, so we're still waiting to find out how the OS stacks up. However, the interim analysis basically showed that there was some improvement in progression-free survival (PFS). It's a very similar study; it's the same chemotherapy, but instead of [adding] pembrolizumab, they used atezolizumab (Tecentriq). PFS was better, but it wasn't “blown away,” and thus far OS has not been any different. That is what is concerning—that we're not seeing any difference in OS. Therefore, we are waiting to see how the OS data matures. Perhaps we will start seeing separations in the curve, but so far, the general consensus in the field is that this is a bit disappointing.

Where do we stand in terms of biomarkers for immune response?

That's kind of the “golden question” right now. I’ll use PD-L1 [expression] as an example, as it's the most tried and tested one. There are so many different ways that it has been evaluated across the board. It does help predict some of those who will benefit [from immunotherapy], but probably the most important thing is that it doesn't always rule out if someone will benefit. Sometimes people have no PD-L1 expression in their tumor, and they may still have a great benefit from [this approach]. Therefore, there’s a huge interest now in determining the best biomarker. There is a great debate, but what's very clear is that we don't have an amazing biomarker yet. That’s why there is so much interest around the world in trying to find one.

What are some emerging markers that are under investigation?

There are quite a few being investigated. There is a lot of interest and debate regarding tumor mutational burden (TMB).

In squamous cell carcinoma, these tumors are [typically in] smokers; it's extremely rare to see someone who gets lung squamous carcinoma and has never smoked. When you look at the counterpart—adenocarcinoma—about 20% of all lung adenocarcinomas are from people who never smoke or are very light smokers. If you look at TMB in those tumors, it is extremely low. We already have pretty good data suggesting that it's about one-tenth the amount [of TMB] as those who get lung adenocarcinoma who do smoke. We already know that the patients who never smoked and get lung cancer don't respond to immunotherapies.

These biomarkers that people are using, such as TMB, are kind of selecting for the same populations. When [a patient is] TMB-low, we now know that corresponds with poor response to immunotherapy. However, a lot of these TMB-low patients tend to also be those who never smoked with lung cancers that have EGFR, ALK, or ROS1 alterations. Therefore, some of the emerging biomarkers are already selecting for things we already knew. That’s why there's some debate about TMB, and about PD-L1 expression and how to test for PD-L1; you can talk all day about that alone.

In terms of other emerging biomarkers, I am certain that there will be quite a few coming. There are many different groups publishing small cohorts of patients, and finding signals that imply that they're going to find something that's better. I have no doubt that with so many people around the world trying hard to find a biomarker, people are going to find better ones. I have little doubt that eventually we're going to figure this out.

Where should future research focus in terms of combinations?

There is so much focus around the world on combinations. [Future research should focus on] logical combinations; there are many combinations that are completely empiric. I've heard that there's actually several thousand clinical trial combinations going on around the world. Almost any kind of permutation you can imagine of different drugs that companies have in their pipeline are being tested; that's completely empiric.

One huge direction that will be important is to have better models. Many of these different combinations of drugs are being tested in people, because we don't have good animal models to test them in. Many of the animal models [used in research and development] have been in lung adenocarcinoma; lung squamous carcinoma is very different. Having good models to be able to evaluate these drugs before we give them to humans is important.

In our laboratory, we were evaluating some models that we made of lung squamous carcinoma. A couple of the combinations that a company had were, to us, very logical. We tried them and it turns out the mice actually survived much less, not longer. That surprised us to the point that we didn't believe the results. We repeated it over and over and kept on getting the same result. This is a combination that eventually we chose not to take to clinical trial because we thought it was actually going to hurt people. It took about 2 years to understand what was happening, but now we actually understand why it would have actually made people's lives end faster, not [help them] live longer.

That is just one anecdote, but it's one example about how when you start giving medicines in combination empirically you can actually hurt people, not help them. As such, the field needs better models to evaluate these combinations before taking them to all of these patients. Because there are all of these thousands of trials going on, we're going to see what comes out of it. Hopefully, we'll see some really exciting results and there will be new treatments.

What kinds of combinations should be explored further?

Many combinations are very centric on immune checkpoint inhibitors and there are just so many different types being tested. However, there are other cells called myeloid populations of immune cells that have different types of macrophages and monocytes; there are a lot of emerging data to suggest that these are bad cells.

Therefore, in terms of the immune checkpoints, we're trying to turn on T cells, which we think are good cells that are attacking the tumor, but we also know that there's a whole other dark side to the immune system. We know tumors [try to use the] myeloid population to their own benefit. The tumors actually want these cells and they're immune cells. Therefore, new therapies that can find ways to block these rogue immune cells may be a very exciting attempt at combinations.

The future is very bright with immunotherapies in terms of combinations; it's going to take the right kind of people thinking about how to combine them and the right ways of testing them. In the end, the proof is in the pudding, which are the patients. I really hope that we'll see evidence that newer therapies will be coming out; it’s all very exciting.