Panelists: Jeffrey S. Weber, MD, PhD, NYU Langone Health; Reinhard G. Dummer, MD, University Hospital of Zurich; Axel Hauschild, MD, PhD, University Hospital Schleswig-Holstein; Michael A. Postow, MD, Memorial Sloan Kettering Cancer Center; Caroline Robert, MD, PhD, Gustave-Roussy
Jeffrey S. Weber, MD, PhD: A burgeoning field is the field of resistance to immunotherapy. Caroline, can you tell us much about what we know about mechanisms of resistance, either primary—meaning to your initial treatment—or acquired resistance after stability or response?
Caroline Robert, MD, PhD: Do you mean to immunotherapy, or to both?
Jeffrey S. Weber, MD, PhD: Yes, usually to PD-1 blockade. That’s where I think most of the work has been done.
Caroline Robert, MD, PhD: It’s true that we should not focus on the underresponders, because we actually have a lot of patients who do not respond and a lot of patients who relapse. So, we are faced with primary or secondary resistance. We have read, in the recent years, some very interesting data on patient samples showing that even in the field of immunotherapy—where until now, we were mostly interested in the microenvironment, the cells that were present here or not, and thinking only about more at the cellular level—we now see, at the genetic level in the tumor cells, we also have genetic events that can actually be the reason for the resistance.
Tony Ribas was really the pioneer to show and publish that some mutations—either in the interferon-gamma signaling pathway, a JAK mutation, some mutation in the beta-2 microglobulin, or in genes coding for protein involved in the immune response at the level of the tumor cells—could be involved in the resistance. So, I think it’s very interesting, because once again it shows us that we should really be keeping our eyes open. We should not assume that we treat with targeted agents so the resistance will be genetic, or that we treat with immunotherapy so the resistance will be to lymphocytes or microenvironment. It’s always the same problem and we have to really take into account the wide spectrum of all resistance mechanisms that can occur.
That said, it’s interesting, but we cannot rely on that for the clinic, because it’s not a very frequent event. I don’t think that if we screen our patients for a JAK mutation we will find. It’s quite rare. It’s just a very big step in understanding the mechanisms of resistance, but it’s not the only one. Plus, we also have read some data in Cell and in other journals saying that a good and functional interferon-gamma signaling pathway could—so the opposite—lead to secondary resistance, because of the emergence of other immune checkpoints that were going to be expressed. So, it’s complex, but we progress.
Jeffrey S. Weber, MD, PhD: What would provoke you, then, to switch therapies? Let’s say you’re giving a PD-1 blocking agent. Michael, what provokes you to switch treatments, and what would you switch to?
Michael A. Postow, MD: If you’re giving a PD-1 blocking agent to start with and a patient progresses, my first question is, do I believe that this progression is an isolated progression? Or, do I believe this is more of a systemic progression? What do I mean by that? If I have a patient on a PD-1 antibody and their tumor burden is generally controlled, but all of a sudden, they have a new lesion pop up somewhere. Sometimes, I’ll think about just surgically resecting that lesion, irradiating that lesion, or treating it with some other type of local treatment. I’m thinking that maybe the PD-1 drug is keeping the whole body in check, but now we just have 1 problem child and we’re going to deal with that with some kind of a local approach.
If, of course, there are multiple areas of progressive disease, then I don’t think that the PD-1 drug effective, and that’s when we would think about drugs like ipilimumab. It remains a question of second-line treatment. Do you give drugs like ipilimumab alone, or do you give a combination: stay on PD-1 and add another drug like ipilimumab to the mix? That is a subject of ongoing randomized trials, so we really don’t know the answer.
Ideally, we would get to a point—as we were talking about with biomarkers—where we could biopsy a progressive lesion, know what the mechanisms of resistances are within that tumor microenvironment, then devise a cocktail for that patient and administer some additional drugs into the cocktail, hopefully to bring them back into response. Maybe you give a PD-1 drug, the cancer progresses, you do a biopsy, and there’s a lot of IDO expression in the tumor microenvironment. You wonder that maybe this is why they’re not benefitting from the PD-1 drug, and then you can add an IDO inhibitor. This is just an example of some thinking in trying to treat patients ongoingly, tweaking the immune therapy as we go based upon different biomarkers either in the tumor microenvironment or in the peripheral blood. We’re not there yet, clinically, unfortunately. We’re still at a point where we’re just trying to understand some of these basic mechanisms. But I think that would be a fantasy world where we ultimately would like to go.
Jeffrey S. Weber, MD, PhD: Would anybody rebiopsy patients to find out those sorts of data and act on them? Or, do you think that’s still in the investigative realm? Reinhard?
Reinhard G. Dummer, MD: I have to admit that we do a lot of biopsies. I think we have to learn, and we try to expand our immunohistochemistry repertoires, or we do PD-ligand-1 testing. We just count the T-cells in the tumor. Sometimes, we do molecular analysis for interferon signaling. We stain for pigmentation genes, proteins, so there are a lot of things ongoing. I think we need to learn. I think the major problem we have is that you need to assess everything simultaneously and put it in large databases and get an algorithm that is clinically useful. And, for this, I think we need close cooperation between us, because nobody can collect the data by their self. We need medical bioinformatic people who can help us deal with the data that we collect.
Jeffrey S. Weber, MD, PhD: Axel, it sounds like some people are doing biopsies, and some people are using their gut feeling. When you have the problem of knowing that a patient has a BRAF mutation, what are the criteria that you use to decide? Are they rigorous criteria, are they seat-of-the-pants criteria, or are they just talking to the patient to decide whether to give them the immunotherapy first or the targeted therapy first? Do you have criteria?
Axel Hauschild, MD, PhD: Yes, before we talk to the patient and discuss the various options, we have a board discussion. So, there are different opinions sometimes on the same patient, whether to prefer tyrosine kinase inhibitors or immunotherapies. This reflects the situation, how biased we are, because there are no clear-cut criteria. The speed of the growth of the metastasis is of importance, there’s no doubt about it. There is the tumor load. We are looking at the LDH, still. It’s a biomarker, and the only one that went to, paradoxically, AJCC classification. We have no other biomarker, and I would love to look in the future at mutational load and other secondary resistance mechanisms.
We have a clinical trial, which is called LOGIC-2. It’s biomarker-driven, so once you have progressive disease, the probes are evaluated and certain patterns are evaluated, and then it’s driven by the mutations that are detected. So, that’s something of great interest. But it’s also the personal situation of the patient. Once a patient is not willing to travel every 2 weeks, every 3 weeks, to our center to be treated, they are treated with oral drugs at home—preferences, and so on. In our center, honestly, 70% of patients prefer immunotherapy once we explain the options. But very many are critically discussing the toxicity of ipilimumab and nivolumab.