Activity Transcript

John M. Kirkwood, MD: Good afternoon, John Kirkwood here, on behalf of myself and Vishal Patel, we will today give a Melanoma Milestones talk, an overview on interesting and I think very current emerging treatment strategies for the disease. I am the Co-Director of the Melanoma and Skin Cancer Program at the University of Pittsburgh UPMC Hillman Cancer Center. Dr Patel is Associate Professor of Dermatology, Director of the Cutaneous Oncology Program and Dermatologic Surgery at George Washington University of School of Medicine. We will overview melanoma and current treatment strategies, the treatment in particular of Stage II and III melanoma as the former have evolved in the past year, clinical trials update, then we will review advanced disease treatment, and then we will turn it to questions.

It is important to mention to everyone that we have a continuously rising incidence of melanoma in this country and worldwide, which has really been the case for a better part of the last 80 years. Here illustrated since 1975 through 2020, where nearly 100,000 new cases are expected this year. Despite this, and I think in part owing to the more than 20 new treatments we have available, mortality has remained remarkably flat over this interval. Melanoma is the fifth most common cancer now. It is projected that by 2040 it may even be the second most common of the solid tumors. And the median age, 65, belies a little bit the statistics because we see a quarter of patients who are under 55 in the prime of their life where it takes a toll, which is second only to testicular cancer in productive life years lost to the disease.

The therapeutic advances we have had include 3 pairs of BRAF/MEK inhibitors, the first being vemurafenib in 2011, followed by dabrafenib, and most recently in 2018 encorafenib. The first MEK inhibitor, trametinib, followed by cobimetinib, then by binimetinib that we will talk about a little later. Equal advances in terms of the immunotherapy of melanoma and other diseases have to do with the recognition of the critical importance of the immune checkpoint, of blockade of inhibitory receptors in the antigen-presenting and T-cell interaction. The first of which to be addressed was the CTLA-4 target, which led to the approval in 2011 of ipilimumab as the first of this class of agent. We have subsequently had anti-PD-1s, and most recently LAG-3, that we will talk about in greater detail later.

PD-1 inhibitors block a more peripheral tolerogenic signal, and has led to, in melanoma, 30% to 40% responses on its own, so they have been widely adopted and widely paired with other agents. The most recent of these is the LAG-3 molecule, one that we have seen the approval of for in March of this year. And I will come back to talk about those later in combination with nivolumab, the agent relatlimab, which is a new option that you all want to be able to consider. With this I will turn the podium to Dr Patel.

Vishal A. Patel, MD: Well, thank you, Dr Kirkwood, and thank you for including me on this presentation. It is an honor to be on the same presentation with you. And thanks for everybody for joining us here on the web. As Dr Kirkwood said, my name is Vishal Patel. I am a dermato-oncologist, trained as a dermatologist, and as a Mohs surgeon, and I focus on cutaneous oncology now, and I head our skin cancer program. I am interested exactly in these types of diseases, treatments that can be treated with surgery but also that require something else, potentially.

I am going to walk us through some of the advances in Stage 2 and Stage 3 melanoma and some of the areas of controversy and maybe the future, to set up Dr Kirkwood to discuss what has occurred on the clinical trials front that has changed the lives of our patients. For those who are not aware, there has been a new staging update in the 8th edition of the American Joint Cancer Commission staging system on melanoma. We have provided that schematic here denoting these relevant changes to the staging system.

Most notably in the staging system, mitosis is no longer a part of the staging and upstaging risk factor. While it is still acknowledged as a high risk factor, it will not change the stage of a tumor solely on that basis. Ulceration continues to upstage tumors that are of the same Breslow's depth. We now see that the decimal points are only to one decimal point with the cutoffs noted here for the earlier thin tumors at 0.8, and then having at every single millimeter difference of 1, 2, and then 2 to 4, and beyond 4 for the relevant pathologic stages. Also, for those who are not familiar, stage 3 disease essentially is any type of nodal disease, either micro or metastatic. You can review some of that detail if you are interested in that. But those comprise the major changes we see in melanoma.

Essentially, stage 2 and stage 3 diseases are thicker melanomas or those melanomas with nodal disease, and how we managed them has changed slightly. I am just going to briefly go through some of those points. The survival curves have also been updated as we have made those slight differences to the categorization of stage IA down to Stage IV melanoma. And we see most notably here how stage III disease on the right has some Kaplan-Meier survival curves that are superior to stage II and IIB and IIC. This has always been the area of most confusion as well as challenges for a clinician to be able to best manage those patients that have metastatic nodal disease, but continue to do better than those patients with non-metastatic disease, but thicker melanomas.

As we will see with Dr Kirkwood's update, we now have multiple options for how to think about treating these, but it can be a difficult discussion for those patients where it is hard for them to understand the dichotomy of not having metastatic disease but being in potentially a worser prognosis group. It is for this reason we have tried to improve on the staging system, and I like to think about staging as more of an analog system that we have utilized with traditional risk factors to compute either with a calculated score or utilizing Kaplan-Meier regression curves, the patient's outcome.

Clearly the future of staging is going to be the genetic expression of each individual tumor and this is the focus of personalized medicine. Melanoma is not unlike other tumors with the use of genetic expression profiling. This can be used for distinguishing whether a nevus is malignant or not, as well as in those equivocal biopsies to get more information to help us denote whether or not we are in a more malignant classification. The area that has seen a lot of discussion for improvement is the area of prediction of risk of a tumor based on the genetic expression. For those who are not familiar with gene expression profiling, there are a number of different tests out there, different companies that market this test. Think of it as a way to assess the expression of potentially bad genes. And can we predict by their variable expression how aggressive the tumor may behave and how a patient may do?

One test that has been produced that is available commercially in the US is a 31-gene prognostic test for melanoma. This test does predict outcomes for disease-free survival, and overall survival of patients with stage I and II melanomas fairly well based on a retrospective cohort. I have noted here some of those tumor markers and genes that they have identified and are able to analyze a specimen and utilize that computational algorithm to classify a tumor as high or low risk.

There have been a number of studies that show that this test is not only able to predict the outcomes of the tumor to the patient, but also the probability of a positive sentinel lymph node biopsy in those lower risk T1 and T2 tumors. As we see here, the take home points I would like to make is that it is clearly notable that the risk of positivity increases as that tumor goes from a class 1 low risk to a class 2 high risk disease. What is important to note is these are retrospective studies. Furthermore, if you look at these percentages, a class 1A result in a patient under the age of 55 suggests a 7.6% risk of sentinel lymph node biopsy. And from the guidelines that we utilize to recommend to patients between a 5% and 10% range is still an area to have a discussion for sentinel lymph node biopsy, and potentially recommend that.

Indiscriminately utilizing this test to just say that we should or should not be performing sentinel lymph node biopsies in a cohort of patients is not quite there yet based on this data. It is notable that Medicare has provided approval for the use of this test for patients who are over 65 with a class 1A result to avoid a sentinel lymph node biopsy in appropriate patients. Again, not using that test indiscriminately. For an older patient who is 80 years old who has multiple comorbidities, it may make sense to utilize this test to avoid a sentinel lymph node biopsy, but I caution against messages that say that this can replace sentinel lymph node biopsy right now.

One of the biggest reasons for that is sentinel lymph node biopsy is still the best tools for determining use of adjuvant therapy, which we will discuss. This test also has a utility in sentinel lymph node negative diseases, and these of course are patients that we worry about that may go on to have poor outcomes. A significant proportion of these patients do have poor outcomes, and this test can help identify those patients that may be at risk for metastases recurrence and potentially death. And we may be identifying how we can utilize this test to determine who we would decide to give adjuvant therapy to.

But, again, critically to note that this is based on retrospective data and as a result there has been a lot of discussion and controversies and concerns about using this in melanoma. I am going to highlight 3 of those that all of you should take home. Because of that retrospective data, these tools are built on those older samples and they lack updated survival information in response to immunotherapy. Furthermore, we do not have prospective trials yet that stratify patients to receive therapy based on the result of this test. Thus, utilizing this in a wide population based way, we are just not there yet.

Furthermore, actionable items, whether that is intervention with adjuvant therapy or just radiological surveillance frequency either increasing or decreasing that are not stratified GEP. When you stratify patients based on lower risk tumor stages, the thinner melanomas, as well as the higher risk stage II and III, there are some differential performances of the test. So while I am a big believer of this test as well as the future of this, the ability to use this widely in every melanoma patient, we are just not quite there yet in my opinion. I think we need to be cautious about that implementation to get that data. That does not mean that this is not going to be the future, I really do believe it will be, but just to think about when you are using that test and what you are going to do with that result. If we do not have that answer, we should maybe think about why we would order it.

Just briefly to discuss surgical treatments, which is the mainstay of stage II and stage III melanoma, has not changed significantly from a guideline standpoint. Recommended clinical margins are noted here and they are variable: in situ disease with between 0.5 to 1 cm; 1 cm for ≤ 1 mm thick melanomas; between 1 and 2 mm thick melanomas there is a range of 1 to 2 cm. And that can be difficult for some clinicians to understand, "Well, should I use 1, 1.5, or 2? And does it make a difference?” Beyond that, we recommend 2 cm clinical margins for > 2 mm thick melanomas. That variability has brought up some discussion certainly in the dermatology world and in dermatologic surgery, and with some surgical oncologists, head and neck surgeons who have trained in Europe and other institutions who are familiar with complete peripheral margin analysis. I thought it was helpful because this has been the update in the surgical realm or at least as an area of discussion and controversy, to understand what is complete margin analysis. How is that different than standard wide-local excision, standard surgical excision, or bread loafing pathology? And is there a benefit between the two?

I think it is notable, especially as a Mohs surgeon, now I might be biased, that understanding what this difference is can really help our patients. The NCCN guidelines, however, note that Mohs surgery is not recommended for the primary treatment of invasive cutaneous melanoma, when standard surgical margins can be obtained. It may be useful in select T1a or thin melanomas in high risk locations, face, ear, or acral sites. But I think what is important to note here is that sometimes we believe that the use of Mohs surgery is synonymous with narrow margins, and that is not the case and should not be the case for melanoma.

Regardless of the technique, it is the margin that matters. It does not matter whether you do standard surgical technique or Mohs surgery, it is that you utilize a standard margin. And surgical margin analysis, whether you use the term Mohs surgery, complete margin analysis, complete circumferential peripheral deep margin analysis. All of these terminologies that are within the NCCN guidelines are descriptive terms for a surgical technique. We have provided 2 images here to show you the difference of conventional excision where tumors are bread loafed or just as it sounds slicing through a specimen like a loaf of bread, and analyze. And then if we look at the whole periphery of the tumors.

I think we can all believe that looking at the whole periphery is better for a disease to check the margin. But the key is that the margin is not shortened as a result of that. You need to still use that standard margin. There are some data that is coming out showing there is improvement of survival in patients, especially in head and neck melanomas, extremity melanomas, by using this complete margin analysis technique. But, again, keeping the margin at that recommended length.

I am going to provide you some examples of what we do at our cancer center. Here is an acral melanoma that required by NCCN guidelines a 1 cm margin. We performed, as you can see here, a kind of hexagonal excision. And each of those specimens, were then process in the Mohs surgery fashion of processing, so we can look at each of those horizontal or linear edges as eight different sections, while still maintaining a 1 cm margin and thus providing a patient with complete margin analysis of that periphery.

For a patient that had a combination of invasive and stage II melanoma on the heel. This also allowed our plastic surgeon to then perform a sentinel lymph node biopsy once we had cleared that, and know the same day that they were ready to have sentinel lymph node biopsy and further the resection of the deep margin all the way down to the appropriate anatomic plane.

Again, another example here, partnership with our ENT colleagues for a large invasive nodular melanoma with a component of superficial spreading. We were able to clear that more difficult component again utilizing at minimum a 1 cm margin around that area to confirm that margin was clear. Then leaving an area that can be easily resected with a 1 or 1.5 cm margin, and a sentinel lymph node biopsy done later that day in the operating room.

Here is another example of a deeper, thicker melanoma, but in an area where a 1.5 or 2 cm was not able to be obtained, and we were able to do that with 1 cm, remove that down, confirmed that was completely clear. While the patient, all these patients have been injected with radiocolloid dye before their surgery, and so they were having their sentinel lymph node biopsy done shortly after this procedure. And final reconstruction, all done in one day in a multidisciplinary fashion.

And so that just leaves me at our last point here that why do we have a multidisciplinary approach? It is because what Dr Kirkwood is going to talk about is that this is now a multifaceted disease. Sentinel lymph node biopsy and surgery is part of the surgical oncologist and dermatologic surgeons purview. But now we have the options of adding therapies much earlier on and not waiting until the advance point where patients are not surgical candidates. And with that point I'd like to turn this back over to Dr Kirkwood. Thank you.

Dr Kirkwood: Wonderful talk Vishal. I think it really set us up for the discussion which will happen in 2 modules. First the adjuvant treatment, which is one that focused upon back in the early 1980s and made the first advances in which is now still evolving. Then we will talk about advanced disease where the progress has been equally amazing, I think. As you will see shortly, allow us to consider newer therapies that may be effective as the older but less toxic.

The first trial we will talk about is KEYNOTE-716, which led to the approval of anti-PD-1 therapy with pembrolizumab in Stage IIB and IIC melanoma, now a year ago. And basically this trial randomized patients 1:1 to receive pembrolizumab or placebo. It was a bomb-proof trial, where we could not fool ourselves by doing observation groups as we used to do in the past. The trial had 2 portions, part one, the initial randomized outcome, where at relapse a patient would be able to be randomized to receive crossover therapy. We will not talk about that today, because the data are really not mature yet for that.

Patients were stratified by T stage and by age. The data as reported already a year and a half ago, where the first interim analysis showed at a median follow up of 14.3 months, a significant reduction in the number of relapses. These data have only matured and become more, I think, salient here at 20.9 months of follow-up on the second interim analysis. And you can look at each of the forest plot categories of patient attributes and disease attributes. The remarkable thing is there was benefit right across the set of facets of disease that are illustrated here, and that really all patients in the Stage IIB and IIC category benefit from this intervention.

In terms of patterns of first recurrence, one of the things that may give us a guidepost to ultimate outcome in terms of overall survival is distant metastasis-free survival. The number of distant metastases plotted here in this pie chart for the pembrolizumab, significantly less, half what they were in the patients who received placebo, giving us hope that survival will ultimately play out. Although it is way too early to have survival data from this kind of a trial, this is certainly suggestive that we will see this.

Distant metastasis-free survival was analyzed by Dr Georgina Long at ASCO the past year. And here you see for pembrolizumab and placebo that there was a significant difference in distant metastasis-free survival. This difference, with a hazard ratio of 0.64, is very close to what we saw for the stage III trials and for the KEYNOTE 716 overall recurrence-free survival. But, again, this is just early, this is just an advance installment on what we hope will be the survival impact.

The downside is there are toxicities. Those toxicities are what we know very well since our first scuffle with CTLA-4 blockade in the 2011 and prior window. Skin reactions, one of the most frequent and often over grade 3, so that they really require therapy. Endocrinopathy with hypophysitis, adrenal insufficiency again seen. But grade 3 or greater only 1% to 2% of patients. This is, while it is notable and it is something we need to present to our patients, it is one which I think when we consider the bigger picture wind up being grade 3 or higher in 10% or less of patients. And something that is I think easily balanced against the risk of relapse and death from this disease.

The CHECKMATE-76K trial compares adjuvant nivolumab vs placebo. The data has not yet been presented, but I think we will likely soon see an additional option for our patients in stage IIB and IIC. Look for the hazard ratio, because again that speaks to the impact but this is really what is still to play out in the next week for us.

COLUMBUS AD is the other option, targeted therapy. We know already that dabrafenib and trametinib in stage III disease has every bit as good an impact as anti-PD-1. And we have yet to see a stage II trial with BRAF inhibitors and MEK inhibitors. So this one will be something also to keep your eyes on, with the combination of encorafenib and binimetinib in patients with stage II melanoma. But this is not yet available to review.

What about unresectable and metastatic disease? This is an area, where as medical oncologists we have spent much of our career. The COLUMBUS 5-year update is I think gratifying here. Dr Reinhard Dummer has presented this and published it this year, and this of course compared encorafenib plus binimetinib to encorafenib alone and to vemurafenib as the reference point, the first of the BRAF inhibitors in play.

And what you see here is that there is clear superiority of the encorafenib plus binimetinib arm to vemurafenib alone. And some benefit over encorafenib alone, which we have seen as well in the comparison of dabrafenib and trametinib to dabrafenib alone, and of course, cobimetinib vs vemurafenib alone. Overall survival here for the same 3 interventions favor the double therapy, and provide now at 6 years follow up a mature basis for recommendations to our patients with BRAF-mutated melanoma.

Grade 3/4 toxicities were very similar with encorafenib plus binimetinib vs the vemurafenib and the encorafenib alone arms. Dose adjustments and dose discontinuations were required only in a small fraction of patients, due to adverse events with this newest of the pairs of BRAF/MEK inhibitors.

There are several trials evaluating BRAF/MEK inhibitors with a follow-up at five years. The COMBI-d with dabrafenib and trametinib show median progression-free survival of 11.1 months and a median overall survival over 2 years. For coBRIM, which evaluated the vemurafenib plus cobimetinib pair, median progression-free survival was 12.6 months, and median overall survival was 22.5 months. For encorafenib plus binimetinib in COLUMBUS, median progression-free survival was 14.9 months and median overall survival 33.6. But these are all in different trials. I think cross trial comparisons we all need to take twice about, these are not compared in any head-to-head setting and they probably never will be.

What about the alternatives? CheckMate 067 is a trial of double checkpoint blockade immunotherapy now at 6.5 years of follow up presented by Jedd Wolchok. This trial compared iplimumab plus nivolumab vs nivolumab vs iplimumab in what has become a real landmark trial. The benefits for iplimumab plus nivolumab are remarkable in that at 5, and 6, and 6.5 years we have half of patients still alive, who only 10 years ago would have been likely to be dead at less than a year. So this is I think a true benchmark. This is the most mature data we have for double checkpoint blockade. This is the reference standard in dual immune-oncology interventions.

The safety here with iplimumab plus nivolumab on the left side with the majority of patients with some toxicity, but with nearly two-thirds who had grade 3/4 toxicity is part of the hesitation that I think people have in the community for applying iplimumab plus nivolumab in patients. But we now have clear data that says this has survival advantages over both nivolumab alone and iplimumab alone, and therefore needs to be our standard of reference.

RELATIVITY-047 is a trial that we were delighted to see 2 years ago at ASCO. The combination of relatlimab and nivolumab, the first of the LAG-3 inhibitory receptor targeted approaches, compared with nivolumab. And the primary endpoint was progression-free survival by blinded, independent review committee. Secondary endpoints included overall survival response rate. And this trial in the intention-to-treat population showed an unequivocal benefit of the combination of relatlimab and nivolumab over nivolumab alone. The hazard ratio here shows a very clear superiority for the combination over nivolumab alone at 0.75.

Not much difference between the ≥1% and <1% ≥1% and <1% LAG-3 expression groups, which in patients who had recurrence after PD-1 interventions was substantially different. Here as upfront, first line therapy, it looks like both groups of benefit, with the benefit a little bit better for the patients with LAG-3 levels that are elevated ≥1%.

The overall survival differences show a trend clearly in favor of the dual intervention, but the time is still early. The hazard ratio of 0.8, is I think something that we will want to follow and keep our eyes on in the future. This newest dual intervention is something that I think we all will have as a recourse for patients who say the toxicity for iplimumab plus nivolumab is unacceptable. Here the toxicity for the relatlimab and nivolumab combination, illustrated on the left, shows you that the toxicities are significantly less than for the dual iplimumab plus nivolumab combination. Rash again plays a part, diarrhea and colitis, occasionally, hepatitis, but overall vastly lower numbers for this combination than what we had seen previously and substantially similar to the nivolumab alone arm.

The question is where does the relatlimab and nivolumab combination lie in terms of our current treatment placements, and our algorithms for treatment? Clearly this is an option for patients who do not want to accept the iplimumab plus nivolumab option.

What does further analysis show across clinical subgroups? Something we will have to look forward to. What is the intracranial activity for relatlimab and nivolumab? What we know is that there is clear activity for iplimumab plus nivolumab. We do not yet know this for relatlimab and nivolumab. Also, what are the roles of biomarkers in this group that might ultimately allow us to select patients selectively for one or the other approach?

The final thing to talk about is that we have finally solved a question, and I have to tell you, Mike Atkins is owed a debt of gratitude from all of us in the field for having stayed this out. This trial, called DREAMseq, was written in 2013 and took us better than 6 years to accrue the desired nearly 300 patients. And the question was, in BRAF-mutated metastatic melanoma, do we start with iplimumab plus nivolumab or do we start with a BRAF/MEK combination? Nobody knew. As you all probably know in the community, the vast preference was for BRAF/MEK inhibitors; they are oral, they do not require visits for intravenous therapy, and the toxicities are modifiable by dose modification in the outpatient setting. So, this trial was really important to us. It randomized patients to receive iplimumab plus nivolumab and at progression dabrafenib plus trametinib, or to receive dabrafenib plus trametinib and at progression iplimumab plus nivolumab. The structure for crossover required confirmation and a bunch of things that would not necessarily delay crossover therapy in standard of practice therapy, but this is clearly something that we need to think about.

Bottom line, landmark survival at both 2 years and at 3 years show a significant difference in favor of the iplimumab plus nivolumab initial therapy followed by dabrafenib plus trametinib, over the dabrafenib plus trametinib in the first line followed by iplimumab plus nivolumab. The benefits were seen in patients regardless of the BRAF mutations, whether they had V600E or V600K. And in each of the subgroups that were analyzed, this difference in favor of the iplimumab plus nivolumab was seen.

The difference was due to a difference in the duration of response. This is illustrated with the iplimumab plus nivolumab response in black and the dabrafenib plus trametinib response in red. Whereas you know and we have already said the duration of benefit from dabrafenib plus trametinib was projected from the literature to be about 12 months, so this is the basis for the difference. The quality of life, both in terms of sleep disturbances and performance and physical function differences all came together by 6 months. I think we can say that support for toxicity from either of these interventions is something we are used to now in the community.

The conclusions of this trial are that iplimumab plus nivolumab followed by dabrafenib plus trametinib is associated with greater overall survival at both 2 and 3 years over the converse sequence. Crossover was frequently not feasible for patients on the trial, where I think they would be feasible in standard of practice application of these 2 therapies. Meaning that if you are treated with iplimumab plus nivolumab , and you had toxicity and you wanted to go to dabrafenib plus trametinib, you would not need to do all of the things that patients on the trial had to do. While quality of life was initially worse with iplimumab plus nivolumab due to the known immune-related adverse events, these all mitigated by 6 months and were similar between the 2 arms. Iplimumab plus nivolumab followed by BRAF/MEK inhibitor, if it is needed, should be the preferred treatment in the community going forward.

To conclude, I think we know that nodal staging, as Dr Patel has presented, remains paramount and is what we should all do in patients with a Breslow thickness over 0.8 mm with sentinel node mapping. Gene expression profiles of several types are available and are going to become part of the algorithm for the future, but right now are still in development. Potentially improved outcomes with complete surgically analyzed margins are something that we hope we can see invoked in the future, and may actually pertain to things like melanoma in situ, early disease, even more than more deeply invasive disease. The checkpoint blockade with anti-PD-1 is effective in node-negative stage IIB and IIC disease, and in node-positive stage III disease, with pembrolizumab approved and likely a second agent soon to become available. In metastatic disease we have anti-LAG/PD-1 combinations that are clearly superior to anti-PD-1 alone, and should displace the consideration of single agent anti-PD-1, I think, in most patients. But then in people with BRAF-mutated metastatic melanoma, CTLA-4 blockade and PD-1 blockade are what should be the standard of care going forward.

This transcript has been edited for clarity.