Dr Igor Puzanov Discusses the State of Adjuvant Therapy for Advanced Melanoma

Igor Puzanov, MD, is clinical professor of medicine at the Jacobs School of Medicine and Biomedical Sciences, University of Buffalo, and professor of medicine and director of the Early Phase Clinical Trials Program, Department of Medicine, at Roswell Park Comprehensive Cancer Center.

In this interview with The American Journal of Managed Care^® (AJMC^®), he discusses the state of treatment selection in the setting of resected advanced melanoma, why sequencing of therapies does not occur in the setting of adjuvant therapy, and best practices to keep in mind for patient education and managing their treatment-related toxicities.

The interview has been edited for clarity.

AJMC^®: What factors do you evaluate when considering the initiation of adjuvant therapy for a patient with resected advanced disease?

We do consider staging of the patient. We have proven adjuvant therapies in stage 2B, 2C, 3A, 3B, 3C, as well as stage 4 NED [no evidence of disease], which means somebody with limited stage 4, which was resected fully with surgery. I’ll take these one-by-one.

Stage 2B, these are the patients with deep primary tumor, either 2 to 4 mm with ulceration—that’s called T3B—or more than 4 mm deep without ulceration, and that’s still 2B. More than 4 mm with ulceration, that’s stage 2C. There was KEYNOTE-716, randomizing patients to a year of placebo or pembrolizumab, which showed improvement in disease-free survival, meaning that fewer patients in the group that received pembrolizumab developed metastatic disease, especially distant metastatic disease. We do not have overall survival data, and that is actually one important thing missing from all these adjuvant trials.

For stage 3A, 3B, and 3C, we also have alternative treatment in targeted agents. If the patient has a BRAF mutation, dabrafenib and trametinib have also showed disease-free survival benefit. There is a whole host of consideration whether targeted or immunotherapy should be used in these patients.

Then for stage 3B and 3C, we have pembrolizumab and nivolumab available. And for stage 4 NED, we have nivolumab and also ipilimumab/nivolumab from the IMMUNED German study as a possibility. So, the considerations are multiple.

One thing I would say about adjuvant therapy: Although it’s FDA approved and disease-free survival has been improved, we still are lacking data from randomized trials between placebo and an active agent, such as the pembrolizumab trial with the EORTC [European Organisation for Research and Treatment of Cancer]. Why is that important? Because patients who are seeking adjuvant therapy come in 3 distinct groups.

One group is actually cured, but they don’t know it and the physician doesn’t know it. These patients will have been overtreated. They simply will have adverse effects, but no benefit really. They just simply get adverse effects. The other group is what I call doomed. They have such a bad biology of disease that whether you give them a targeted agent or immunotherapy, they simply will progress and die because the tumor is not accessible to the mechanism of action of the drugs we have available. Those patients are either undertreated or overtreated because they need something different, better—but what is that? Then, of course, what you are giving them is not beneficial. Again, they will have adverse effects and no benefit. And then there is a third group, which is destined to develop metastatic disease. And it’s destined to develop disease that will be sensitive to immunotherapy with checkpoints or targeted agents. But what we don’t know about those is that you simply cure them, but [would you have cured] them if you detected them when they developed an early metastatic disease.

We know that with immunotherapies, we now actually have a median overall survival with ipilimumab/nivolumab of 72.5 months. With single-agent PD-1 [programmed cell death protein-1], it’s less, but there are still long-term survivors. And that’s the problem with adjuvant therapy for me.

At the end of the day, you are simply not sure whether you are curing the people you would cure if you just waited long enough when they develop metastatic disease and you overtreat or undertreat the other 2 groups. You have to weigh it together with potential for adverse effects, where the checkpoint inhibitors will carry a risk of death. For single-agent PD-1 agents, whether it’s nivolumab or pembrolizumab, that’s 0.4%, mostly from cardiomyositis [myocarditis]. And there is about a 10% risk of severe adverse effects, which require hospitalization, and about 40% [have] long-term sequela—chronic side effects such as arthritis, weakness, multiple endocrinology problems, which stay with the patients—and although they are easily to be remitted, they still carry some risk.

If somebody develops adrenal insufficiency, you can put them on a long-term steroid supplement. But if they get in a car crash and people don’t realize they have adrenal insufficiency, they may die from adrenal crisis unless they get the stress dose steroids. Remember, this can be the same patient who was destined to just survive with the surgery alone.

All of these problems are artificial and unnecessary, but we don’t have a system where we could actually select them properly. We really need data from randomized placebo vs active agent trials, such as the latest in the stage 2 or the one I mentioned, the EORTC in stage 3. What this gives you is a look at patients who were randomized on day 1, but also patients when they developed distant metastatic disease.

It would be really interesting to see overall survival data from the trial, because if the overall survival from the trial is equal, then I would argue that you can simply wait with immunotherapy, screen patients for a metastatic disease, and treat them when you detect one. Of course, opinions differ. And I talk with the patients about these pros and cons, the relapse-free survival vs adverse effects and potential for death with immunotherapies or adverse effects with targeted agents. The benefit of targeted agents is, of course, that there is no risk of death. They don’t kill people. The adverse effects go away quicker than with immunotherapy. But at the same time, they’re still there.

We don’t have data from a targeted agent vs immunotherapy adjuvant therapy trial. If the overall survival data for placebo vs an active agent actually show the survival benefit, then I would say that then it’s clear that you have to treat everybody, overtreatment or not. And we need to work on biomarkers to hopefully select patients better, either up front or after a couple doses of the immunotherapy, to actually see who really benefits and who either doesn’t or maybe is harmed.

AJMC^®: How do you sequence adjuvant therapy options? How do BRAF mutations impact treatment selection?

We don’t sequence adjuvant therapy options; we simply select one or the other, or none. It depends on the BRAF mutation status. We check BRAF mutation status in all patients with stage 3 and higher disease. It is unclear if it’s better, for the appropriate stage patients—which really would be stage 3A, 3B and 3C—to do immunotherapy or targeted agent. Because selection of the agent may affect effectiveness of the drugs in the stage 4 setting.

But stage 3A actually has a sufficiently low risk of progression and death, that if a patient has a BRAF mutation, I tend to use a targeted agent in patients with stage 3A disease. In those with 3B, 3C disease, I’m more on the side of immunotherapy. However, after discussions with patients, we personalize to the patient’s preferences, as well as other more subtle considerations. For example, somebody with stage 3C disease who really wants to delay the onset of the disease, at the same time, they suffer from rheumatoid arthritis—I would give them the targeted agent. Otherwise, I may select immunotherapy for that patient with stage 3C disease.

AJMC^®: How will the discovery of tumor mutational burden, a possible predictive biomarker, impact management of care? Could it enable us managing melanoma going forward?

Not at all. Tumor mutational burden is really not a predictive biomarker for anything, with any strength of evidence. We really don’t have any biomarker.

AJMC^®: What are some best practices that we should keep in mind when we're managing the adverse events that patients experience?

There are actually good guidelines for managing checkpoint-associated immunotoxicity. I would mention the SITC [Society for Immunotherapy of Cancer] guidelines, the effort I originally led, and we now have a second edition; the ASCO [American Society of Clinical Oncology] guidelines; the NCCN [National Comprehensive Cancer Network] guidelines. We published a book with SITC on management.

What is clear for managing checkpoint inhibitor toxicity is that the education of both physician and of patient is of utmost importance, and good communication, quick communication when the adverse effects potentially appear. Knowledge of the communicator on the physician’s side is important, so they can tease out from the patient the appropriate information. For example, a patient reports loose stool. You need to know how they look. Are they watery, is there blood, is there pain, is there fever? How many a day above a normal level? And then you have to refer to the CTCAE [Common Terminology Criteria for Adverse Events] toxicity criteria and appropriately grade it, because if you don’t appropriately grade it, you don’t know what treatment you should choose and you don’t know what you should do with the offending agent (ie, stop it; treat and restart; or stop it, treat, and don’t start). That is really important, and people are still not doing that.

It is mind boggling, because we have electronic everything. There is UpToDate, there is Google, there are guidelines online, and people simply do not use that, which is surprising to me. The training, I guess, has changed, and people simply do not read anymore and do not know how to interpret the data and how to listen to the patient. But that’s the fact and you have to adjust to it.

It’s absolutely clear that you have to talk to the patient and inform them before you start so they can make informed decisions. Give them written information. I use UpToDate or other patient-related education materials; they are in writing. We give laminated cards with the name of the drug, name of the physician, phone number to call during hours, after hours. So if a patient travels, they can call 24/7 somebody who is knowledgeable about their situation or their doctor who is treating them.

Where there is maybe controversy, but I don’t see controversy myself, we actually really worry about cardiomyositis, because although it’s rare—0.5% in a population—it is about 30% to 50% fatal. This is actually how people die with immunotherapy; it is usually cardiomyositis, and result in hard blocks and arrhythmias. Few patients still die from pneumonitis and few from colitis, but that should be eliminated hopefully soon. But cardiomyositis is different because it can come between day 7 and day 60, with a peak between days 15 and 42. So we actually check baseline troponin and EKG on every patient starting the checkpoint and then weekly troponin alone for the first 6 weeks, because it's usually all or nothing and comes very early.

One thing to remember when patients get immune-mediated adverse effects: They may not just get one. They may actually get multiple. For example, you may have cardiomyositis and you may also have liver toxicity. At the same time, you could have rhabdomyolysis that actually comes together, muscle and muscle damage. On the other hand, it’s treated with steroids, so you get it both. But sometimes people will develop endocrinopathies, like hypothyroidism, hypo-adrenal state, hypophysitis. One has to be on a lookout for either concurrent adverse effects or sequential. You may recover from one adverse effect and 2 months later, you may have another adverse effect without any more treatment with checkpoint inhibitors. The cells just get angry with other organs. That’s important to be on a lookout for. About 30% of patients will have multiple, either sequentially or at the same time.

The pills, they are easier. The one thing about pills, you stop them, it goes away. With the MEK inhibitors, such as trametinib—because we talk about adjuvant therapy and dabrafenib—trametinib is the only FDA-approved adjuvant therapy with targeted agents. It’s not encorafenib and imatinib. It’s not vemurafenib, cobimetinib; just dabrafenib/trametinib, and that’s what I use.

You have to watch for depressed left ventricle systolic function with serial echoes, like before and every 6 months or so. You have to watch for arthralgias, myalgias, pain in the joints, muscles. Fevers can come, and they have to be managed—that’s mostly from the dabrafenib portion. And again, the MEK inhibitor may cause serious retinitis, which means a patient develops blurry vision. So you tell them to stop the pills, go to see the eye doctor. They look, and they see fluid behind the eye that usually goes away without treatment. You can even restart the pills. Sometimes they have some rare averse effects, like damage to the neurons, inflammation; we’ve seen a few cases of Bell palsy, but it’s rare and usually goes away with stopping the therapy. What we do, and again, there is no randomized data for that. Although, we sent an abstract to ASCO.

When we start the pills, we start at a lower dose, at 50% dose reduction. And we have printed documents for the patient that if they tolerate the 50% dose reduction—we call it dose level 1—after 5 days when they reach a steady state level with these pills, they go to dose level 2, 3, 4. So, within 2 weeks, we have them actually on a full dose if they tolerate it. The difference is that instead of starting a full dose, if they develop these adverse effects like fevers, having to backtrack, we simply ease them into it from the bottom up and we can actually stop the adverse effects before they get too severe.

But again, with a nonrandomized and somewhat selected population, it’s hard to say that this is the right way for everyone. But we do, and it’s actually really easy. One side effect of that approach, you actually realize which patient is more compliant than the other patient. Because if they can follow the schema of the dose escalation, they are telling you that they pay attention to adverse effects and are good communicators. The ones who don’t, you realize this could be a problem, because a patient is having hard time managing the adverse effects, the pills, and you may actually devote more resources to that patient to help them navigate through the treatment. Because they may need help with other pills. We always look at other medications and whether their medication regimen can be simplified. We work with the primary physician, cardiologist, [and] others.

AJMC^®: What are you working on in advanced melanoma that you are excited to share? What’s next on the horizon for you?

We don’t typically switch adjuvant therapy. Remember, we start adjuvant therapy within 3 months from the surgery. So, let’s say a patient starts, 1 month or 2 months out from surgery, dabrafenib, trametinib, and they develop fevers. We try to adjust doses. We are now 4 months in, 5 months in, and they still cannot tolerate it. I’m not going to give them nivolumab or pembrolizumab to make up for it; we are done. In the same way, if they don’t tolerate the immunotherapy, they develop adverse effects, we have to stop it. I’m not going to put them on pills, because I have no data on a sandwich managing of adjuvant therapy. There is no switching. You select one and you go 1 year or whatever time, and once you finish, you are finished and you go on surveillance.

In advanced melanoma, so we have several types of patients who are not being served well at this time. We are working on the primary resistance to immunotherapy. That’s the patient who actually comes in with either BRAF mutation or BRAF wild type nonmutated and they start immunotherapy, because we now know that starting with immunotherapy saves more lives than starting with targeted therapy. So regardless of BRAF status, they get immunotherapy. And they progress within 3 months; they never had a response; they just simply go through it and there is nothing. Those patients die very quickly; there’s no really good salvage treatment for those patients.

We have a Department of Defense supported trial, where we use a dendritic cell vaccine combined with a mixture of cytokines, which stimulates the innate immune system like toll-like receptor agonist 3/4. We also use interferon in that mixture of cytokines. So that’s one.

Also, there is a new potential breakthrough treatment on the horizon, which is now in a randomized phase 3 trial. It’s called IO102/IO103. These are actually 2 vaccines, peptide vaccines, which target cells making either PD-L1 [programmed death-ligand 1], which would be a tumor, or IDO [iondoleamine 2,3-dioxygenase], which is the enzyme associated with the immune-suppressive environment in the tumor, where the tumor tries to protect itself from the attack by T cells. And you combine them, they’re actually pretty nontoxic, with pembrolizumab.

There is a randomized trial, either pembrolizumab or the mixture of the 2 vaccines and pembrolizumab, and the beauty is that we are making the memory T cells, which actually remember the IDO and they remember the PD-L1. In the nonrandomized phase 1B trial, the response rate was as high as 92% if the PD-L1 of the tumor was more than 1%. It was a small data set, about 53 patients. It was nonrandomized.

The next step we want to do is to combine with combination immunotherapies, because nowadays, the nivolumab, ipilimumab, as well as the LAG-3 inhibitor, relatlimab, seem to be taking off as the way to use immunotherapy in melanoma. That’s what we are working on right now.