In this interview, Dr. Michael Postow, Assistant Attending Physician with the Melanoma–Sarcoma Oncology Service at Memorial Sloan-Kettering Cancer Center, and Dr. Tony Nimeh, Research Physician at Brigham and Women’s Hospital, discuss advances in immunotherapy for melanoma.

Dr. Postow is enthusiastic about the promising early data on the effects of blocking the programmed death-1 (PD-1) receptor using nivolumab and lambrolizumab, antibodies with specific activity against PD-1. However, he cautions that therapy using immunomodulators needs to be administered with care because of immune-mediated side effects and the occasional need for immunosuppressants. Large phase III studies will be necessary to determine the full potential benefits of these PD-1 immunomodulatory drugs. The ultimate hope is to win FDA approval for the drugs, making them accessible to the many patients in need.

Dr. Nimeh: Dr. Postow, would you discuss the recent data on nivolumab that were presented at ASCO?

Dr. Postow: Nivolumab is one of several very promising new antibodies that target the PD-1 receptor. Previous data on nivolumab had been published in 2012 in The New England Journal of Medicine¹ This year at ASCO, long-term data from those studies were presented, and we learned more about PD-L1 as a potential biomarker for PD-1 therapy with nivolumab.

Dr. Mario Sznol presented long-term data on PD-1 treatment with nivolumab in the Melanoma/Skin Cancers Oral Abstract session.² The drug toxicity profile was extremely favorable, even over time and with continued dosing. The lack of accumulation of additional toxicities over time is obviously very important with any ongoing, continuous therapy.

The median overall survival also appeared quite favorable, when compared with prior historical data in advanced melanoma. The median overall survival was 16.8 months across all of the different doses that were evaluated. Of course, cross-trial comparisons need to be interpreted with caution, particularly in a phase I trial such as this one; but, these results are certainly promising and, based on the results, several larger-scale, randomized, phase III studies are being conducted to test the drug’s benefits in a larger population.

Potential biomarkers for Nivolumab

Dr. Nimeh: Are there any immunotherapy biomarkers that might help identify the best candidates for nivolumab?

Dr. Postow: At this point, there are no specific biomarkers that have been studied thoroughly enough, and provided data that are convincing enough, to influence the selection of a patient for PD-1 antibody therapy. PD-L1, however, is being examined as a potential biomarker for nivolumab in ongoing large phase III trials.

To put this in context, we need to keep in mind certain principles in immunotherapy. Most anticancer therapies target the tumor itself. With biomarkers, we attempt to detect genetic changes in the tumor, or changes within oncogenic signaling pathways, to help us direct therapies at the tumors. In contrast, all of these immunomodulating antibodies (such as PD-1) act on, and target, the patient’s normal immune cells rather than the tumor itself. Because of that difference, it has been very difficult to identify a biomarker that would influence patient selection. In concept, every patient could potentially benefit from immunotherapy if it was used appropriately and in the right sequence or combination.

There is a lot of discussion about PD-L1 as a potential biomarker for nivolumab and perhaps other PD-L agents. In biomarker studies of PD-L1, patients are considered positive if they have at least 5% membranous staining on the tumor surface by immunohistochemical analysis. These patients preliminarily appear to have a greater likelihood of tumor shrinkage. However, some patients have an excellent response even though their tumors do not have PD-L1 expression. Therefore, patients who do not express PD-L1 should still be considered for treatment with agents such as nivolumab.

Nearly 50% of patients with melanoma have a BRAF mutation. Patients with BRAF mutations are expected to do just as well with nivolumab and other immunotherapies as patients without BRAF mutations. There are no convincing data to suggest that the response to immunotherapy depends on the particular mutation found in the tumor, especially with BRAF. Often, small-molecule targeted therapies, such as vemurafenib, dabrafenib, and trametinib, are approved by the FDA and used for patients with BRAF mutations. Combinations of these targeted agents are also being tested to see if they are better than single agents themselves. 

The specific sequences or combinations of therapies that are most beneficial are still being evaluated in clinical trials.  For a patient who has a mutation such as BRAF, we do not yet know, based on available data, whether it is preferable to start treatment with immunotherapy or with targeted therapy. It has been our practice, for an asymptomatic patient (who does not need an immediate response), to start with immunotherapy. This gives the patient an opportunity for long-term durable benefit. However, for the patient who has symptomatic disease or a large tumor burden, we prefer to begin with targeted therapy, as we feel this gives him or her the best chance for immediate tumor shrinkage (which seems to occur more readily with the targeted agents).

Dr. Nimeh: Blockade of the PD-1 and PD-L1 pathway may be effective against tumors in which PD-L1 is expressed at the cell surface, reflecting a baseline, “preactivated” state of the immune system. Is there anything we can do to up-regulate PD-L1 in cells that are not in such a preactivated state?

Dr. Postow: That’s an excellent question and one that is the focus of several ongoing investigations. It is not yet completely clear which specific treatment modalities would up-regulate PD-L1 because we have not yet done enough prospective trials, with pre- and post-treatment biopsies, to assess PD-L1 changes associated with a specific intervening therapy.

PD-L1 is believed to be up-regulated in the setting of chronic inflammation. This may be a mechanism by which the tumor cells create their own resistance to otherwise effective antitumor immunity. When tumor-infiltrating lymphocytes interact with the tumor, it has been shown (through interferon-gamma secretion) that PD-L1 can be expressed. Thus, perhaps the tumor has found a way to protect itself from these immunologic onslaughts of tumor-infiltrating lymphocytes. Perhaps we need to find ways to increase this immune-cell infiltration within the tumor to cause a PD-L1 up-regulation leading to an increased susceptibility to the beneficial effects of PD-1 therapy.

Bear in mind that we have not yet determined the appropriate sequence or combination to up-regulate PD-L1 clinically, after which we could use the PD-1 antibody to enact a response. It is a critical point that even patients whose tumors are PD-L1 negative can still respond to PD-1 blockade.

Side effects of immunotherapy

Dr. Nimeh: What are the side effects of immunotherapy to be aware of, and how do you manage them?

Dr. Postow: Immunotherapy enhances the body’s immune system, and our hope is that this results in improved antitumor immunity against the malignancy. Unfortunately, this activation of the immune system can also result in reactivity against tissues not involved in the malignancy, creating important side effects. And since patients often require immunosuppression for the treatment of immune-related side effects, the potential for the development of opportunistic infections is important to keep in mind.

Inflammatory pneumonitis is particularly relevant for antibodies that block the PD-1 axis. A few patients, unfortunately, died from complications of pneumonitis in one of the phase I trials. However, with increasing recognition and prompt treatment using immunosuppressants, we believe that this potential side effect can be effectively managed. The number of patients who experience pneumonitis is very low, particularly those in whom it occurs to any significant degree.

Renal insufficiency is another side effect—elevated creatinine levels have been associated with the PD-1 agents. Colitis and diarrhea seem to be more of a problem with CTLA-4 blockade than with PD-1 blocking agents.

The main point to remember, though, is that all of these agents can cause immune-related adverse events, which are generally easily managed with prompt recognition and the initiation of steroids or other immunosuppressants. Long-term sequelae are generally extremely mild, if they occur at all.

No maximum tolerated dose has been established in dose-escalation studies with PD-1 agents. These drugs are very well tolerated because the mechanism for adverse effects is different from that which occurs in chemotherapy or targeted therapy.   

Measuring response to immunotherapy

Dr. Nimeh: How is response to immunotherapy measured, and how is “success” defined?

Dr. Postow: Ultimately, we will assess the response to immunotherapy differently from the methods we have been using up to this point. Right now, however, response to immunotherapy is assessed using traditional radiographic response criteria, such as those applied to other investigational agents. Basically, these are the modified World Health Organization response criteria or the RECIST criteria.

Traditional response endpoints, apparent on CT scans or PET scans, can often be misleading and potentially underestimate the long-term survival benefits accrued with immunotherapies. The reason is that the patterns of response that occur with immunotherapy are different from those observed with traditional chemotherapy. For example, with immunotherapy, new lesions may appear in the presence of an otherwise decreasing tumor burden, and there may even be an apparent worsening of overall tumor burden, prior to ultimate shrinkage of the tumor burden and a response.

The immune-related response criteria have been investigated for patients treated with ipilimumab (an anti−CTLA-4 blocking antibody now approved by the FDA), and it was proposed that a novel set of response criteria should be applied to that drug. In fact, the novel immune-related response criteria appeared to better approximate the long-term survival benefits of the drug as compared with traditional radiographic response endpoints. The newly proposed response criteria still need to undergo prospective evaluation. Nevertheless, immunotherapies, particularly the PD-1 agents, demonstrate high response rates even with traditional radiographic response criteria, such as RECIST.

We are still struggling to establish the best surrogates to apply in early-phase trials and to ascertain how well these surrogates correlate with overall survival. Overall survival, therefore, remains the most important endpoint in clinical trials evaluating the immunologic agents. Interestingly, a greater number of patients have achieved long-term survival after treatment with ipilimumab than were considered to have responded by traditional radiographic response criteria.

Treatment regimens in immunotherapy

Dr. Nimeh: With new treatment approaches emerging, what is the potential for combining immunotherapies—for example, ipilimumab and granulocyte-macrophage colony-stimulating factor (GM-CSF)—or adding targeted therapy?

Dr. Postow: Many new agents appear to be promising as monotherapies. It is very exciting to think about the possibility of additive (or potentially synergistic) benefits if we combine more than one in appropriate and rational ways.

At ASCO this year, Dr. Stephen Hodi presented data which showed that patients who received ipilimumab 10 mg/kg in combination with GM-CSF had improved overall survival as compared with patients who received ipilimumab 10 mg/kg alone.³ The results were consistent with those from preclinical evaluations in mice. This combination treatment demonstrates the potential promise of combining these immunologic agents.

We reported in The New England Journal of Medicine,⁴ and also presented at ASCO this year,⁵ that, in a phase I trial involving a small number of patients, the combination of nivolumab with ipilimumab resulted in a very impressive response rate.

Whether or not these responses to combination nivolumab and ipilimumab (which have been durable) translate into long-term overall survival benefit is the subject of an ongoing phase III study. At this juncture, we do not yet know whether it is better to combine nivolumab and ipilimumab initially or to sequence them, one followed by the other. Patients can respond to nivolumab even if they have not responded to ipilimumab and vice-versa. We are hopeful that the large-scale phase III trials will answer these questions regarding the relative merits of combination therapy vs monotherapy. Combinations of targeted therapy and immunotherapy are also very exciting to investigate. Several prior evaluations^6,7 have suggested that treatments such as BRAF inhibitors can result in immunologic effects that may be additive or synergistic when combined with immunotherapy.

We conducted a phase I trial in which we combined vemurafenib with ipilimumab.⁸ Unfortunately, the trial was discontinued because the rate of hepatotoxicity was too high. One of the many lessons we learned from that trial was the importance of carefully evaluating combinations of targeted therapy with immunotherapy in early-phase clinical trials, in which dosing and scheduling are very carefully monitored. It is possible that sequencing these agents, or using specific combinations, will be more tolerable and possibly beneficial.

One ongoing study is evaluating the use of dabrafenib and ipilimumab, with or without the MEK inhibitor trametinib.⁹ We are hopeful that this study will provide some additional information on the possibility of combining targeted agents with immunotherapy for patients with advanced melanoma. At this point, we would not recommend combining immunotherapy and targeted therapy outside of the context of a clinical trial. We encourage the enrollment of patients in such trials.