Melanoma Podcast

Survivorship, Surveillance, and How Long to Follow a Melanoma Patient Population

Sapna P. Patel, MD; James Larkin, PhD, FRCP, F Med Sci


September 19, 2023

This transcript has been edited for clarity. For more episodes, download the Medscape app or subscribe to the podcast on Apple Podcasts, Spotify, or your preferred podcast provider.

Sapna P. Patel, MD: Hello. I'm Dr Sapna Patel. Welcome to Medscape's InDiscussion series on melanoma. This is episode 5. Today we'll be discussing the duration of follow-up for different melanoma populations. First let me introduce my guest, Professor James Larkin of the Department of Medicine at Royal Marsden Hospital in London. Welcome to InDiscussion, James.

James Larkin, PhD: Thank you very much for inviting me. It's great to be here.

Patel: We've been doing an icebreaker question for all of our guests. For our listeners, is there maybe some fun fact or something that you can share that people may not already know about you?

Larkin: Some people won't know that I've got a fitness coaching qualification — CrossFit coaching — which I got a few years ago in the US, because there was a program for MDs to get a CrossFit Level 1 certificate. So I went to California and I did the course and I got it, but I'm not actually doing any coaching. What time I have, I concentrate on trying to get fitter myself.

Patel: Maybe at one of these medical meetings, you can put us through a boot camp. Maybe as you do for your team at retreats.

Larkin: Maybe. There's a few people in the melanoma world who are kind of big CrossFit people. Allison Betof Warner is one of them. So maybe that's a serious proposition.

Patel: Let's talk about surveillance in melanoma populations. There's obviously surveillance for patients who have had their primary melanoma resected. We have what feels like fairly arbitrary rules through the National Comprehensive Cancer Network (NCCN), and maybe the American Cancer Society, about how long you're supposed to follow patients after a primary tumor. What is your take? Are we seeing, in the big adjuvant trials that have been done, that 5 years is sufficient follow-up to find these metastases and then it sort of plateaus? Or do you find that longer follow-up detects more metastases?

Larkin: That's a great question. The thing is, we've got a few different populations, and if we look at adjuvant trials — let's say with anti–PD-1, or dabrafeniband trametinib — we've got about 5 years of follow-up or just over. We can look at the shape of those curves, and clearly they're flattening by the 5-year time point. And in fact, they flatten quite early. But for me, the big question is what happens after 5 years, and I think the reality for that at the moment is we don't really have data, number one. And then number two, I'm not sure you can assume anything in the absence of data about what might happen, let's say, between years 5 and 10.

And just to throw something else in there, I don't think you can assume that different drugs with different mechanisms of action will necessarily follow the same types of patterns, and you can broaden that out a little bit as well, if you want to. Then the tradeoff beyond 5 years is going to be, if you do follow-up beyond 5 years, when do you stop and how long do you do scans for? It's got a lot of different implications, not least how the patient views their situation and the whole use of the term "cure." So I'm not really sure at the moment for me anyway what to do for adjuvant treatment. I'm mainly talking about stage III disease here, bearing in mind that we've had more recent data for high-risk stage II disease on exactly what we should do. I've tended to be conservative in clinic, or what I think is conservative; this is usually to do a set of scans at year 7 and a set of scans at year 10. A set of scans means thorax, abdomen, and pelvis and a brain scan. It would also probably be reasonable to do annual scans per year between years 5 and 10. At the end of the day, we just don't have the data, but I hope over the next little while we'll get some data so that they can inform these conversations with patients and guidelines.

The final thing to say is that resource utilization differs in different parts of the world. So, depending on where you practice, there will be pressure on clinics. There will be pressure on scans. There will be reimbursement issues. Obviously that shouldn't be first and foremost, but I think it is something that needs to get taken into account as well.

Patel: I think there's a couple of things there. We will only get these data if we in those large adjuvant trials continue to get the sites covered to see these patients and enter the data. I hear this talked about sometimes in the renal world. If you look farther out on a Kaplan-Meier curve, it's fewer patients at risk. What do you do with that information? You don't have as much population at risk. You have maybe more censoring due to loss to follow-up. So what can you really conclude from what might be 10% of your population who still have data or things being entered past 5 years? Those adjuvant trials may not have really budgeted to do this long-term question, but it would really behoove us to make sure we're doing that.

We finished a SWOG trial called S1404 a few years ago. It's actually the largest adjuvant anti PD-1 trial completed. It had 1300 patients, and we're probably past the 5-year mark now, but what we'll probably have to do is a strong push of all of those sites to say, "Do you still have patients who are being followed, and if they are being followed, can you try to capture their data?" In SWOG studies in the US, we say that you have to be followed for 10 years and after 5 years, it's just annual scans like you were saying. So we might have something from there. I don't know so much about the targeted therapies or some of the pharma-sponsored trials, if they wrote to follow patients long-term like that.

But neither you nor I can admit to be health economics experts. What does it mean when you're ordering scans on even 10% of your population to pick up whatever percentage we might be saying — for example, 1%-3% of those late scans? What are we saying in a privately insured population vs a nationally insured population? You start to wonder whether there's actually dollars and cents that go behind this long-term follow-up.

Larkin: I agree with all of that. The thing is, if you're picking up a late melanoma relapse, this is still probably someone who is potentially curable of the disease. If we're talking about health economics and you can then successfully treat the disease in the long term, that's obviously of value to the patient, that goes without saying, but actually there's an upside there from a resource perspective as well. On the other hand, you don't want to be following up a population where your event rates are really low, but I guess we still don't know what the event rate is at the moment.

A couple of other points that obviously in the industry-sponsored studies, it's in their gift how long they follow up the studies, give or take, whereas at least with an academic study like SWOG, you have the power as an academic group, assuming that the funding is available, to follow it up. That's pretty important for an academic group. That's a really worthwhile use of funding, because the question is going to be there until it's satisfactorily answered.

A few other things to bring up are is issues related to long-term side effects and survivorship. I still think that in terms of data, there's some pretty big gaps. Again, an academic group is going to be well positioned to try and answer some of those questions because if we think in the even longer term — let's say cardiac risk or cardiovascular risk or vascular risk, these kinds of things — people talk about this a little bit with checkpoint inhibitors and probably targeted therapy as well, but do we really know whether there's a change in cardiovascular risk if we've cured someone's melanoma?

The other thing that you imply from what you said was disease-specific survival, which is obviously critical. We have some data in melanoma for disease-specific survival in CheckMate 067, but the disease-specific survival data are pretty scant, other than those in a kind of systematic way, as far as I'm aware. In melanoma, the patients are younger and generally less often have vascular problems than do patients with kidney cancer or lung cancer, and there's a really big question for checkpoint inhibitors, adjuvant and metastatic, about disease-specific survival in some of these other diseases. That's an important gap as well.

Patel: Let me circle back to that disease-specific or melanoma-specific survival in a second. As we're talking about long-term surveillance and long-term toxicity, it makes me think about these survivorship clinics that we've developed, at least in the US. They're trying to say that past a certain number of years off of treatment in cancer A, B, or C, you're now considered for a survivorship path. The National Cancer Institute says you're a survivor from the day that you're diagnosed, and although we agree with that, the question becomes, who is to follow for these long-term toxicities? Is it going to be in the hands of the primary care physician? The cardiologist? Should it be the oncologist who's following these patients long-term and trying to unpack, "Well, you're now hitting the age range of cardiovascular disease, but we also gave you a checkpoint inhibitor for a couple of years…" Who should do some of that risk unpacking? In the US, our survivorship clinics probably aren't even structured in a way that can dive deep on some of these things. Where you practice in the UK, do you have survivorship paths or pathways for patients to go on to once they've finished their melanoma treatment? Or do they stay with the oncologist? Do they go back to the GP? How does that work?

Larkin: In a word, the answer is no, we don't have any organized survivorship clinics. Patients would go back to their GP, they'll probably be having dermatologic surveillance, but that is going to largely be for new primary disease and nonmelanoma skin cancer. So we have a gap there. It's something we're trying to address at the moment in an academic way, in a research-type endeavor, but until we know more about the long-term side effects of these drugs, we absolutely should be doing this.

The other issue is, and you said it really, that most oncologists aren't trained to do any of this stuff. So you need a collaborative type of approach; maybe the cardiologist is the most important person to follow someone. But at the end of the day, things are set up in such a way that although ideally it would be multidisciplinary, probably it's going to need to come under the auspices of someone. Then I'm guessing that in the US, you've got reimbursement and other challenges, unless it's clear that there's a particular need for this. And we're still at a stage where we'd be looking to do it with charitable funding, that kind of thing, rather than being in a place where we can set up a routine clinic in the way that you would for survivors of childhood leukemia or hematologic malignancies or other childhood malignancies.

Patel: You can feel that it will become quite a popular topic or specialty — a post–checkpoint inhibitor surveillance or survivorship clinic. As we're using these agents now across so many solid tumors, you would think that the kidney cancer patients and the lung cancer patients and the melanoma patients should be followed with some sort of structure, because the side effects they might have long-term may be more unique to the treatment than to their cancer or necessarily their demographic.

In the adjuvant setting, we think about how long to follow patients up. Maybe we'll say 5 years is as much data as we know, but we all know that we have cases of patients who experience recurrence later. Fortunately for us, like you said, most of those recurrences might still be in a curative phase because isolated brain metastasis, or brain metastasis in addition to extracranial, wouldn't be the most common scenario for presentation. It would be mostly extracranial metastases. If that's the issue — if we're following now to pick up these specific melanoma recurrences but we still look at metrics for federal agencies (such as the US Food and Drug Administration and the European Medicines Agency) and we're still looking at things like overall survival (OS), now we're having a harder time teasing apart this younger melanoma population that ages and may not actually be dying of their melanoma; they could be dying of something else. What, then, does OS mean when we're treating earlier and earlier stage 2 and stage 3 disease? Is OS really what we should be capturing, or as we design the next batch of clinical trials, do we need to start telling the sponsors that it has to be disease-specific? We really care if patients are succumbing to their melanoma and it's maybe a secondary measure, competing risk estimands if they're succumbing to something else. I think it will help us from getting confused by lack of OS benefit when the deaths and the things that we may see with OS may not even be related to patients' melanoma. What do you think?

Larkin: That's quite an interesting point. I don't know if we're ready yet to make that transition in terms of a primary endpoint for trials. One potential disadvantage would be that we've got a vast amount of experience of trials with OS to compare with in terms of benchmarks, but that isn't to say it isn't important to collect disease-specific survival. I don't know enough about the hematologic literature, leukemia and lymphoma, to know.

In other words, they cure lots of patients as well, and I'm sure they haven't abandoned OS, but I don't know whether or not they've made any kind of transition to disease-specific survival or not. That's the thing that immediately came to my mind. This isn't straightforward, particularly if you're thinking about registrational studies which are designed to get drugs approved and reimbursed. Changing the endpoint for those kinds of trials practically might not be straightforward. But there isn't anything to stop you, if you're designing a trial, to try as much as is possible to have this as a key secondary endpoint.

Patel: As you did in CheckMate 067; as you said, there was presentation of disease-specific survival, which did have more attrition and had more censoring, of course, because when you don't know what somebody died of, you really can't call it necessarily a melanoma-specific death. But I think that was a start in seeing that.

First of all, the shape of those curves per my memory looked the same — melanoma-specific survival and OS — and in a metastatic population, that's the lowest-hanging fruit. In these earlier settings where we might have cured patients, that's where it's really hard to tease apart. First of all, how long do you have to follow? Then when somebody dies 10 or 20 years later, how do you access that to see whether it's related to or unrelated to their cancer? Maybe it does behoove us to look at some of those more highly cured cancers and see how they determine survival.

That makes me want to transition to the neoadjuvant setting, because now we're thinking about giving therapy before primary surgery. As we presented event-free survival for S1801, there was a lot of, "Oh, but the OS was not significant." Of course, the OS was immature. We hadn't yet hit a median number of events. But we give a lot of adjuvant therapy in the absence of OS data. So do we need to be hanging our hat in the neoadjuvant space on OS if it doesn't change what we do in the adjuvant setting? That's number one, and number two, is this another scenario where maybe the next neoadjuvant trials should have this surrogate of melanoma-specific survival — are we doing something melanoma-specific to improve patients' survival? My concern is that as we start throwing more and more combinations in the neoadjuvant space, is some of our survival compromised because of toxicity? Is it possible we're going to have some toxicity-associated demise? That's a little harder to quantify, but if we actually try to separate OS from melanoma-specific survival, maybe we could make a start on this.

Larkin: I don't think that I have any good answers to any of those questions. Let's take adjuvant therapy to start off with — as you say, we've been given quite a lot of adjuvant treatment in the absence of OS benefit. Do we think there's an OS benefit for our adjuvant treatments? I guess the answer is yes, and there's been some indirect modeling. I don't think it's plausible, really, with checkpoints inhibitors in the adjuvant setting that there isn't a survival benefit.

But then next up, will it be the same with neoadjuvant therapy? I think we need to wait and see, honestly. If you take S1801, there's relatively limited follow-up; as you said, inevitably, people will be interested in survival in due course. I'm sure there will be a survival benefit, but we need to see that.

And then your point about designing the next set of studies: This is a nice opportunity to really get this on the agenda because people are going to talk about it. These are studies that we'll often report relatively quickly, in comparison with some of the studies that we do. So, I'd be pretty keen on that.

Part of having this discussion and doing things like this means that down the line, when people are talking about this and reviewing the manuscripts and everything else, there's a real discussion in the field and more broadly in oncology as well about this. I'm pretty positive and supportive about that.

Patel: Back to adjuvant S1404, we reported OS when we reported the relapse-free survival (RFS). And it was mainly because we did a timed OS. We said X number of years after the last patient is enrolled, we'll do the one-time look at OS. This was to avoid dragging out a federally funded trial for years and years to look at the survival.

It ended up being negative — no statistical significance in OS. It's the only adjuvant checkpoint inhibitor trial that's looked at OS, but we also didn't have a median number of survival events there because we did this timed OS look. When do you think we'll see anything with CheckMate 238 for an OS analysis?

Larkin: That's a good question. The problem there, obviously, in CheckMate 238 the active control arm is ipilimumab (IPI), which in fact did have an OS benefit in the original European Organisation for Research and Treatment of Cancer (EORTC) trial, and there's been an indirect comparison between the two arms for OS. So again, biologically, I can't really see how you wouldn't have an OS benefit, but it might not be that we can demonstrate it.

Then you slightly come back to the "now or later" discussion as well. In other words, if we've got potentially curative treatments for stage 4 disease, should we just wait and not treat any early melanoma and wait for the ones who do develop metastatic disease and treat them then? This is obviously an important question, for all sorts of different reasons. But, I would mention that in the EORTC adjuvant pembrolizumab study, Alexander Eggermont, Caroline Robert, and others tried to address this now or later question, prospectively and formally. I don't know what you think about the data, but they're quite difficult to interpret, honestly, with 5 years of follow-up.

Part of the problem with pembrolizumab for stage 3 disease vs pembrolizumab for stage 4 disease, which is the trial design basically, is that if you've got someone who's presented with stage 4 disease and they haven't had adjuvant treatment, you might not automatically want to give them pembrolizumab. It could be years later, the standard of care could have changed, there's all sorts of variables there. It's inherently difficult and time-consuming to do this. This debate rages in kidney cancer now, where pembro has an adjuvant label globally, there's distant metastasis-free survival (DFMS) benefit in kidney cancer, the trial's not mature for OS, and lots of people are saying, "There's no OS benefit, we're not going to give the treatment." It's interesting in melanoma, we never really had that particularly. You could speculate on why there's that difference — but on the other hand, if you've got a patient in clinic and you've got a solid trial and you've got a label and you've got a treatment that, in my view, at least is likely in kidney cancer to result in an OS benefit but the trial's just not mature yet, are you going to say now I'm not going to treat you, I'm going to wait for 2 or 3 years?

That's the kind of tension in the system and the tension in the consultation room as well. It's trite and cliche, but I guess there aren't really any easy answers, unfortunately, in the absence of usable markers, which we haven't really talked about at all.

Patel: You said a few things there that are key to say. CheckMate 238 and S1404 had an active control arm. So if we actually don't see an OS benefit, it may be because we're not talking about doing something vs doing nothing. We're talking about doing something A vs doing something B in which you might only see an RFS, not OS. That then means that maybe we should be waiting for something like KEYNOTE-054 (the trial by EORTC and Alexander Eggermont you were talking about), or even KEYNOTE-716 (the stage 2 pembro vs placebo trial). These trials that cross over and allow for treatment with pembro at either recurrence, after you've done pembro, or crossover from the placebo and pembro at your early stage 4 disease maybe are the trials from which, when they read out, we'll be able to figure something out.

Progression-free survival 2 can be a very confusing endpoint you were alluding to, and people have a hard time knowing what to do with that. You and I both have been here long enough to remember the adjuvant interferon days, and there was a real conversation on the lack of OS — why would you give such a toxic regimen, et cetera, et cetera. Well, we still have lack of OS in melanoma, but is it that our regimens got easier and we're happy to use easy treatment still in the absence of OS? It's interesting that the denominator hasn't really changed that there is no OS benefit, but we definitely swung the pendulum from very few people giving adjuvant interferon — and I happen to be one of them if they had an ulcerated primary tumor — to a lot of people giving adjuvant checkpoint inhibitor.

Maybe at the end of all of this is that patient perspective. Because you and I can sit and think about OS and DMFS and RFS, and ultimately what we hear all the patient advocacy groups say, and what a patient will say in the consultation room, is, "I just don't want this to come back." Of course, they want to put their brains around living longer, period, but they don't even want that next recurrence; they want all of this in the rearview and they don't want to be thinking about more melanoma treatment in 2 years or 5 years. So RFS and DMFS are incredibly important and emotional endpoints for a patient. Regulatory agencies have a lot of weight they want to put on OS. So there's a disconnect already there between what is important to a patient and what is important to a regulatory agency, don't you think?

Larkin: I couldn't agree more. Our patients are exactly the same, and we're 4000 or 5000 miles away. You sit with a patient — and I've probably had this conversation hundreds of times over the years — and it's very rare for someone not to want to have a well-tolerated treatment. I would make that point with a large effect size, and I'm comparing interferon now, which we know in the metastatic setting is associated with an OS benefit, so it would be a bit odd if it wasn't in stage 2 or stage 3 as well.

The other thing from the patient perspective: Let's say we weren't giving any adjuvant treatment for stage 2 or stage 3 disease. We all know that with the best will in the world, sometimes people are going to have relapse with untreatable or very difficult-to-treat stage 4 disease. That has haunted me, honestly, since the start of my career, that whole question. Of course, we can't identify the patients and all that sort of thing, but to have missed the opportunity to give a treatment that may have stopped that happening on an individual level feels really difficult as well. So the patient perspective is absolutely critical here.

Patel: This has been a great conversation. Maybe we didn't change any of our unknowns here about how long to follow patients — we still probably don't know how long to follow-up an early-stage patient — but it feels like we agree that 5 years feels arbitrary. There are recurrences that happen past there. I like your strategy of maybe year 7, maybe year 10, and maybe in some situations, even annually, between years 5 and 10. But ultimately, let's collect some of those data and see how long to follow these patients. I think with neoadjuvant therapy, we will have to report OS. The paper by Reinhard Dummer on neoadjuvant T-VEC vs surgery just came out saying that at 5 years, there is an OS benefit. So if there's efficacy and an event-free survival, there should theoretically be some signal in OS, but it becomes hard as you lose patients to follow-up and maybe they die of other things, competing illnesses, and not actually their melanoma; that can also impact that endpoint.

We didn't really touch on how long to follow a patient with metastatic disease, but that's more challenging. I think your CheckMate 067 trial really has changed our landscape. We've gone upside down now, from maybe 20% of metastatic patients living 5 years to maybe 80% living 5 years with metastatic melanoma. In that case, that probably tells us we have to follow those patients now even longer than 5 years, because something can happen with that stage 4 no evidence of disease status. But we love the tail of the curve that you see developing at 3 years on CheckMate 067, so that's fairly reassuring. I'll pause here and see if you have any other final thoughts on surveillance.

Larkin: Like you said, we had a nice discussion, but I'm not sure we've answered the questions. The European Society for Medical Oncology (ESMO) meeting is coming up, and I think we're going get a bit more data beyond 5 years in the adjuvant setting, and I'm sure there'll be some discussions at the time. So people need to watch that space.

Patel: Sounds good. Today we've talked with Professor James Larkin about survivorship, surveillance, how long to follow a melanoma patient population, and as you heard, maybe more data emerging at some upcoming conferences.

Thank you all for tuning in. If you haven't done so already, please take a moment to download the Medscape mobile app to listen to and subscribe to this podcast series on melanoma. This is Dr Sapna Patel for InDiscussion.


Malignant Melanoma

NCCN Guidelines. Melanoma: Cutaneous Melanoma

Malignant Melanoma Staging

A Practical Guide to Understanding Kaplan-Meier Curves

SWOG Cancer Research Network

Adjuvant Pembrolizumab Versus IFNα2b or Ipilimumab in Resected High-Risk Melanoma

Long-Term Outcomes With Nivolumab Plus Ipilimumab or Nivolumab Alone Versus Ipilimumab in Patients With Advanced Melanoma

Estimands — A Basic Element for Clinical Trials

Neoadjuvant-Adjuvant or Adjuvant-Only Pembrolizumab in Advanced Melanoma

Definitions of Additional Oncology Drug Endpoints

Adjuvant Nivolumab Versus Ipilimumab in Resected Stage IIIB-C and Stage IV Melanoma (CheckMate 238): 4-Year Results From a Multicentre, Double-Blind, Randomised, Controlled, Phase 3 Trial

Adjuvant Ipilimumab Versus Placebo After Complete Resection of High-Risk Stage III Melanoma (EORTC 18071): A Randomised, Double-Blind, Phase 3 Trial

Adjuvant Pembrolizumab Versus Placebo in Resected Stage III Melanoma (EORTC 1325-MG/KEYNOTE-054): Distant Metastasis-Free Survival Results From a Double-Blind, Randomised, Controlled, Phase 3 Trial

Pembrolizumab Versus Placebo as Adjuvant Therapy in Completely Resected Stage IIB or IIC Melanoma (KEYNOTE-716): A Randomised, Double-Blind, Phase 3 Trial

The Validity of Progression-Free Survival 2 as a Surrogate Trial End Point for Overall Survival

Neoadjuvant Talimogene Laherparepvec Plus Surgery Versus Surgery Alone for Resectable Stage IIIB-IVM1a Melanoma: A Randomized, Open-Label, Phase 2 Trial

Follow Medscape on Facebook, X (formerly known as Twitter), Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.