Optimizing Immunotherapy Remains a Challenge in NSCLC

David R. Gandara, MD

Immunotherapy is rapidly altering the treatment landscape in non—small cell lung cancer (NSCLC), but many questions still remain, according to David R. Gandara, MD.

OncLive: What is your reaction to the MYSTIC trial findings?

In an interview with OncLive, Gandara, director of the Thoracic Oncology Program, professor at UC Davis Comprehensive Cancer Center, and a 2017 Giant of Cancer Care in Lung Cancer, discussed the latest immunotherapy developments in lung cancer and the next steps toward optimizing the benefit of these agents.Gandara: The results of the MYSTIC trial have been much anticipated. This is a phase III trial in advanced-stage NSCLC, randomizing patients to standard-of-care platinum chemotherapy versus either durvalumab (Imfinzi) alone or durvalumab plus tremelimumab, a CTLA-4 inhibitor.

In a press release from AstraZeneca, it was reported that the primary endpoint of progression-free survival (PFS) was not met in either of the experimental arms. This is only a press release, so no data were reported in terms of the magnitude of benefit; it was just reported that the primary endpoint was not met.

This could still play out in multiple ways. The study will not be presented until the 2017 ESMO Annual Congress in early September. It is possible that, with further follow-up, there could be improved overall survival even though the PFS was not improved. It could also be that the magnitude of benefit is different between the 2 experimental arms, or that the degree of PD-L1 expression, which was at the 1% or greater level, could differentiate patients who benefitted versus those who did not. I want to hear the complete data set from this trial. We have to reconcile it against the other studies.

If we put MYSTIC together with the negative results of the CheckMate-026 nivolumab (Opdivo) trial, published by Dr David Carbone and colleagues in the New England Journal of Medicine, it means that trials of checkpoint immunotherapy are not always going to results in a big advance. It’s not guaranteed that it’s going to hit a home run in every setting, and we’re still very naïve in our ability to design trials that will carve out the patients who will benefit the most.

Those negative results also emphasize how impressive the data are from the other first-line trial, the KEYNOTE-021 pembrolizumab (Keytruda) trial with a similar design in advanced-stage NSCLC, which demonstrated improved PFS with patients with a high PD-L1 expression of 50% or greater compared with platinum-based chemotherapy. Even though crossover was built into this study, it also showed improved overall survival (OS). This means that the patients on the chemotherapy arm who crossed over to pembrolizumab and had a benefit did not interfere with the ability to show improved overall OS.

How do you think clinical trials like these could be better designed?

Lastly, the recent FDA accelerated approval of pembrolizumab plus platinum-based chemotherapy in all comers without any PD-L1 expression requirement, is in contrast to all of these studies. In view 2 out of 3 first line immunotherapy trials being negative in advanced NSCLC, I am concerned that the accelerated approval may not pan out in the phase III trial in that setting, which is the KEYNOTE-189 study. The data on whether to test for PD-L1 and who to treat based on the results are confusing right now. Both patients and practicing oncologists are going to be confused until the remainder of these studies are reported. How to design clinical trials is an ongoing debate. At one point, 10 years ago, I analyzed the last 24 randomized clinical trials that were done trying to add a targeted agent of many drug classes to chemotherapy versus chemotherapy alone. Of those trials, only 2 were positive in unselected patient populations. The first was the bevacizumab (Avastin) trial from ECOG and the second trial was the partially biomarker-driven FLEX trial of chemotherapy with or without cetuximab (Erbitux).

If you have a targeted drug, it is important to define the population who is most likely to benefit and perhaps design a biomarker-strategy that incorporates only those patients. Therefore, the drug and its companion diagnostic are more likely to have a positive result.

That being said, many of the trials being done today want to throw the drug at every patient with NSCLC, regardless of biomarker status. My own impression is that the checkpoint immunotherapeutics are targeted drugs. Just like for EGFR mutation or ALK rearrangement, we have to find the appropriate biomarkers.

Right now, it’s PD-L1. In the future, it may be an immune signature, tumor mutational burden, or a combination of things; however, until we find better biomarkers for these drugs, we will likely see some positive trials and negative trials without a rhyme or reason why they are so.

Speaking of biomarkers, can you discuss the efficacy of PD-L1 as a biomarker?

Can you discuss how the FDA approval of frontline pembrolizumab has changed the treatment landscape?

For example, nivolumab is a very similar drug to pembrolizumab, so why would one of the first-line trials be positive and the other negative? Are there differences in patient populations, the biomarker cutoff that was used for the trials, or even the assay that was used? We don’t know for sure.PD-L1 as a biomarker has had ups and downs. There is an old analogy about a new drug saying that when it first comes out, everyone thinks it’s fantastic but, after a while, it loses its luster. Two years ago, most people were turned off by PD-L1. We have swung full circle with data of pembrolizumab as a first-line therapy, showing not only an increase in PFS but also OS, despite crossover. This has led to where PD-L1 tested with the Dako IHC 22C3 assay has become the standard of care for first line therapy. Now, with some of the data coming in with chemotherapy plus pembrolizumab, or even the MYSTIC trial, we are still grasping for what is the best biomarker for the checkpoint immunotherapeutics.My own opinion is that single-agent pembrolizumab [is appropriate] for patients with PD-L1 expression greater than 50%—it’s important to define that population. It means that, in the United States, there is a reason to test every patient. Also, if they’re going to be tested, it should be done with the Dako IHC 22C3 assay. If the results are greater than 50% in non-squamous cell stage IV disease then, in my opinion, they should receive pembrolizumab.

There are many caveats. As I mentioned earlier, the recent accelerated approval by the FDA of pembrolizumab with platinum-based chemotherapy with no requirement for PD-L1 expression has left us in a quandary. It’s difficult. You can imagine a patient could be told, “We’re going to test your tumor for PD-L1. If it’s greater than 50%, we’re going to give you a drug called pembrolizumab. If it’s not, we’re still going to give you pembrolizumab but also throw chemotherapy together with it.” Hopefully, as more results come in we will reach some clarity on these points.