Well, Mike and I had the pleasure of leading the group workshop on capitalizing on the promise of immunotherapy.

I would like to start off by thanking Mike for all his help before the meeting and during the meeting. I lost the coin toss.

So, I am up here to try to explain what the summary of our discussion was.

I would like to thank everybody for participating. It was a very active discussion. We covered a lot of issues and I would welcome anybody's help as I go through this, or when we finish, to either mention some things that I either stressed inadequately or over-stressed, and if we have left anything out.

TOP

It started out quite nicely, almost in complete agreement that the main problem we had was that, despite the presence of immune responses to carefully identified, well done, well delineated targets, that tumor regressions are rare. That started out as our main problem.

TOP

We did well in collecting and listing what people agreed was our accepted body of knowledge, starting off with the fact that tumor associated antigens do exist and that actually, if you look hard enough in many patients, particularly melanoma, that you will find immune responses to these tumor-associated antigens.

It was also pointed out that sometimes you find responses to these tumor-associated antigens in normals, patients who don't have melanoma as well.

We also agreed that, without focusing on which particular immunization regimen people might use, that immunization in general, when performed and monitored, would lead to identification of an expansion of tumor reactive T cells, and also, if you look for them, you can see antibody production, suggesting that immunization did lead to enhanced immunity.

TOP

In the past, there was a lot of talk about, how do you monitor. I think there has been a lot of excellent work done, summarized incredibly well by Ulrich and Jeff in a Journal of Immunotherapy review, from a Society of Biological Therapy workshop.

We now have actually a significant number of assays that we can perform that are reliable and can actually tell us about the quantity of T cells that are produced in response to immunization and also, if you do the right tests in combination, you can assess the function of these T cells.

Despite this, we all agreed that, when you look at patients with advanced disease, clinical responses are rare and, with the current state of the knowledge, when you do find clinical responses, that there is really no good correlation between the clinical response and the immune response. That is sort of a major problem with where to go. So, that is the body of knowledge we started with.

TOP

The next area we discussed a lot about, and what I have summarized in just a few bullet points here is, where are the scientific opportunities. What are the scientific questions that are unanswered that can help us solve our main problem?

One of the first issues was immune regulation. Clearly, many of the antigens that we looked at are self antigens. So, there are normal immune mechanisms that ar at work to try to prevent immune responses.

There are now a number of interesting possibilities to look at into how to avoid this immune regulation.

We know that there are CD4, CD25 positive suppressor cells. We know there are potentially suppressive cells in the myeloid series.

We know that there are normal mechanisms to control immune regulation, like up-regulation of CTLA4, and there are antibodies that one can use to inhibit CTLA4.
Interestingly enough, trials with anti-CTLA4 that have been done actually show tumor regression in small numbers of patients treated with the antibody.

In the same venue of immune regulation, there was enhanced activation of T cells. There are co-stimulatory models. 41BB, OX40, it was felt that these were potential areas that need to be investigated to help us solve the main problem.

The other opportunity was felt to be about the cell transfer, a strategy that I sensed in the workshop was resurrected.

We know there has been a lot of work done with adoptive immunotherapy, with NK cells and LAK cells and TILs that really haven't become part of standard care.

However, it is hard to ignore the recent results with using the lymphodepleting regimen followed by adoptive transfer of tills. It led to large numbers of patients experiencing tumor regression, and it was felt that this was an important scientific opportunity and that this work should be continued and duplicated.

There is a lot that we don't know about how it works that would also provide us with new information for better clinical trials to help solve the main problem.

Certainly, related to that, is this whole idea of lymphodepletion and homeostasis, immunological concepts that were not around or poorly understood years ago.

Clearly, lymphodepletion was a part of the adoptive cell transfer experiments that Mark Dudley and Steve Rosenberg did, but it is not clear that you need to do adoptive transfer of tumor infiltrating lymphocytes.

There is good animal data that we discussed that showed that this could be done in a setting of a bone marrow transplant, where non-immune activated cells could be the transfer component.

There are other animal data where just normal spleen cells can be used in combination with vaccination, and it was felt that this represented a good scientific opportunity and clinical trials ought to be performed addressing all the variety of questions raised by those studies.

Another important area is understanding the tumor microenvironment. We talked a lot about the tumor and the idea that the tumor -- people in this room know there is a variety of mechanisms by which the tumor can evade the immune system.

Tumor heterogeneity was discussed, looking at antigens on the tumors, the ability of tumors to actually process and present the appropriate antigens, a lot of sense that you can't just focus on the T cell end of things, that you have got to focus on what the target is doing, and not just because of the antigens it might or might not express, but because of all the things that it might produce, such as the TGF beta, interleukin 10 and the arginine story from Agusto Achoa that actually suggests that you can interfere with T cell function at the site of the tumor.

These were all felt to be poorly understood areas of the immunobiology of the tumor host interaction that need to be addressed and solved, and needed to be addressed not only at the preclinical level, but it really appears to the committee that there were really clinical opportunities in each of these areas.

Of course, cytokines, as you heard before, they are the only approved therapies that we have, other than DTIC, for the treatment of melanoma, and it was felt that the current cytokines we have that do work are worthy of further investigation and that we decided we were up to IL-29 -- that there are other cytokines that are worthy of study in this disease alone, or in combination.

This actually represented the meat of the discussion and these are the scientific opportunities. As you can see from the breadth of these, they don't lend themselves to one or two easy solutions.

TOP

There were a number of areas of controversy, and we talked about targets, trying to tie our session in with the other sessions.

Although people agreed we had targets, there was a significant disagreement as to whether the appropriate targets were, as

Frank suggested in his session, that they were unique targets, they needed to have something to do with the transformed phenotype.

There is a camp obviously that thinks that is the way to go, but there was another group of individuals who felt that shared antigens are what we have, shared antigens are what we should use, and that it is appropriate to study them. So, that was nothing that we were going to resolve in yesterday's discussion.

There was a lot of discussion of the role of immunological monitoring. Okay, we have these tests, they are all better, they are reproducible, but what role do they actually have, and has the advent of these tests actually helped us figure out why some treatments work and others don't, and there was a difference in opinion of the value of creating these what are actually very expensive tools to accompany clinical trials.

There were people who felt that that wasn't an effort that was worthwhile, and there were others who thought that this was still a very important part of what we do, and should be done in all studies, as much as possible, whatever the expense.

Another area of discussion which we may have come down differently than Ulrich and Franco's group -- and you will hear from them next -- was the idea of what do you need to go forward with a new regimen.

Go forward sort of means, when would you be ready to test a regimen in a non-disease setting. There was clearly a group that felt, you have to have responses, you have to go to metastatic disease when -- whatever strategy you are trying, you need to go to patients. You need to have evidence of clinical tumor regression before you move to the earlier setting.

There was another group that felt that if you actually had an active specific immunotherapy regimen that you knew worked, and worked in this case was to induce an immune response in the majority of patients that meets a sort of agreed upon threshold, that that was actually adequate to move to an earlier setting.

There were others who felt that, well, maybe there is an intermediate end point such as stable disease. So, you don't have to get actual tumor regression. Maybe we can look at the timing of prolongation and stabilization of disease, and maybe that is a good enough clinical end point, if we can measure it.

I think this is a reflection of the lack of surrogate end points that help us predict, and one of the clear deficiencies was a lack of biological surrogate end points.

I have to admit that there was clearly a T cell bias in terms of the people invited to that workshop and present at that workshop.

There was a significant discussion about how little T cells have added and how poorly the T cells have worked thus far, and maybe a complete focus on T cell biology is not an appropriate one.

There have been significant advances in our understanding of NK cells, NK cell biology, and how they can mediate antitumor effects.

There was a plea made, and I believe heard, for us not to ignore the role of antibodies in the treatment of this disease, acknowledging the fact that the approved biological therapies that we have out there are monoclonal antibodies.

There was a lot of discussions about, well, the targets on those cells are different and they actually target molecules that cause apoptosis, or they target molecules in Herceptin's case that are signal transduction.

So, there was a lot of discussion about, well, maybe the same kinds of things do exist in melanoma. I think it was an area of controversy. I think the general sense was that we need not be T cell specific in our view of how immunotherapy might be used for the treatment of melanoma.

TOP

Where is our understanding lacking? Where isn't it, is how I would answer it, in our session.

I mean, there is just -- we know a lot more than we used to. We have a lot of opportunities, but there is still just a lot more that we need to learn.

Although the cytokines work, we don't know how they work. It would improve our ability to take the field forward if we knew how interleukin-2 was effective, if we knew how interferon mediated its effect.

There was a sense in that regard that, because of that, we don't really have good selection criteria for patients in clinical trials. I mean, they have to pass the IL-2 stress test, but wouldn't it be nice to have selection criteria that came from microarrays or whatever method, that would allow you to determine who had a 50 percent or 80 percent chance of responding to IL-2.

It was felt that more information could help us make those decisions. Clearly, one of the big problems that remains is how do you immunize patients appropriately and, without going into discussions about DC's and IV and subcutaneous and all these things, there was a clear consensus that we need to understand better how antigen presentation occurs, and what is the optimal way to achieve that.

The idea of trafficking, so much of the immune system has to do with the effector cells getting to the site of tumor, things that we don't understand, not just effectors, but the idea that the stimulators, the APCs, need to get to where these responses occur.

There is a big void in these areas, and they are all fruitful areas for further investigation at both the preclinical and the clinical.

Then a word you heard before, I think I know what it means, but the idea of the plasticity of tumors. They are not this inert object and they are worthy of our study.
If one is only studying the T cell part or the effector part, you are forgetting half the equation and you are not going to be able to solve the problem.

So, there was a real emphasis placed on studying the tumor. Based on this, you can probably predict what some of our recommendations were going to be.

TOP

The major impediment to progress -- these are other impediments to progress: the lack of clinical responses, lack of surrogate markers, and I won't belabor this, but we also had the discussions of the inability to get reagents.

Our discussions, since some of the people who brought that up here were in our session, they centered around the same notion that it would be nice to have a better way of getting the cytokines and other biologics and antibodies made, getting them earlier, and collaboration between different pharmaceutical companies, if you have different cytokines from different people you want to put together. That is just restating what was said earlier.

TOP

So, our solution seemed simple minded and pretty obvious. That is, we need to solve the scientific problems that we identified.

I wish there was like one thing we could say to do that, but I can't -- and I will challenge anybody else that was there -- but I can't escape the notion that we still have a lot of work to do in multiple areas to solve fundamental biological and immunological problems before we can make specific comments about how to solve the problem.

So, our solutions are actually tied to all the problems I raised. There are solutions and there are clinical and preclinical ways to address all those issues.

TOP

In an attempt to provide specific recommendations beyond solutions to the problems we raised, we thought there was value in a retrospective analysis of the archive samples from responders and non-responders on the trials that have already been done on patients who have received and responded to IL-2 and interferon.

As we heard from the meeting parts yesterday, those studies are all underway. Steve Rosenberg and Mike and others have started the exchange of those samples, and I know that John has plans for a lot of the archived samples on interferon patients from the adjuvant trials.

We feel that is a first step of samples we already have that may help us get an answer, or at least get some hypotheses that we can test in other clinical trials.

The other component that you have heard, we obviously need to perform more clinical trials, but we need to perform them carefully in a way in which there are adequate numbers of patients placed on these trials.

There needs to be careful consideration that the appropriate patient materials, whether it is serum, peripheral blood, tumor cells, lymph node samples -- and it may be all of them for some studies, and maybe it should be all of them for as many studies as we can do -- to collect those and to make sure that we are capable of performing a detailed analysis of responders and non-responders as they arise in the studies, testing the variety of different strategies we outlined before.

TOP

There was a suggestion that this was best accomplished in a multi-institutional consortium. You have heard this from every group so far.

Better collaboration, I think, is what people are calling for.

There probably aren't -- all the single institutions can't do all the tests with all the samples.

So, somehow we need to collaborate. We need to work together. Perhaps we need to make cores.

There was discussion about how this would interact with the cooperative groups and the sense that the cooperative groups aren't really ready to handle this kind of strategy currently, and that perhaps this would exist at some level below the cooperative group with intergroup mechanisms, and that this would be a place where new ideas and new translational things would flow from to the cooperative groups.

Then there was an opinion to establish a progress review group in melanoma, and we heard what the director thought of that last night. Okay, thank you.

TOP

MICHAEL ATKINS: You know, in all human tumors, we don't need to announce to this group that the data to support a role of immunity in melanoma is better for any other human tumor.

Yet, our ability to exploit that understanding has been devilish at best. It used to be that we didn't have antigens. We have more than enough antigens. It used to be that we didn't know how to elicit a T cell response. Yet, we have got substantial ability to now both elicit a T cell response and identify it.

I think if I was going to side with any of the lists of problems that you identified, it is getting back to the tumor biology, in other words, understanding the critical role of the interface between -- the stromal-tumor interaction, but for us, it is the immune-tumor interaction, understanding that deeply.

I think the best way to do this is in the clinical materials that we have. Don Morton's concern about melanoma models, I think, is germane. I think many of the things we have to learn in melanoma we can get from our own patients.

I would like to urge this group -- because we have substantial materials already -- that we use information and materials that we already have and study our own patients, using modern technologies.

Rather than picking out the things we care about -- T cells, antibodies, DCs, cytokines, whatever -- that we do broader technology-based screens to identify the next set of questions. The ones that we thought were the problems are not the problems, I don't think.

JEFF SOSMAN: I guess what I am going to say is a little heresy, but the question really is, should we sort of back off immunotherapy.

The pie is limited. I mean, there is a very limited pie, and we know that some of these efforts, the adoptive ones, especially, are incredibly expensive.

I guess, you know, is it worth continuing to put this kind of effort, and should we sort of go back to the drawing board, which could include looking at samples, but could include looking at new animal models.

The human clinical trial experience has been so disheartening, I guess you have got to ask if it is worth putting a lot of effort at this time into it.

PARTICIPANT: As a non-immunologist, let me encourage you to continue these studies, but then stepping back and asking some fundamental questions, particularly, are we really identifying the right tumor antigens.

I am not sure that we have done enough of the basic immunology side, and it could be done potentially with the tissues and going back, which T cells are the most responsible for killing.

The antigens that have been selected over the last few years by the community, I personally am not very convinced that they are the ideal antigens to target.

So, having a discussion on what are the basic mechanisms of immunity and what is going wrong at the moment, I think that would be very timely. My co-chair wants to address that.

DR. NISHIMURA: Actually, I want to address Jeff's comments, and there are two points I would like to make. The first is, adoptive cell transfer is a very powerful technique, not just for treatment, but also for understanding what goes on.

If you put in cells with defined specificities, you can monitor them, you can traffic them, you can understand how these cells behave in ways that you can't do if you just vaccinate a patient, because you don't know necessarily what to look for, and how to follow these cells in vivo.

In a lot of the animal studies with labeled cells, they have gone a long way to teach us what actually happens to cells in vivo.

I firmly believe that these studies will provide us a tremendous amount of information as to how T cells and other effector mechanisms work in humans.

A second point about cost. You know, we have no problems doing a stem cell transplant and generating patients that have chronic GVHD and treating them over courses of years, with very expensive treatments.

There are other treatments that are out there for cancer patients that are incredibly expensive, what it costs to keep HIV patients going with their drug regimens. I think cost is not an issue.

If this actually ends up treating patients and getting responses, then the cost is really not the issue. I would encourage us not to look at financial considerations when we talk about developing therapies that might actually be beneficial.

That is what we were just complaining about with the drug companies, that the drug companies look at the bottom line and they don't push forward with treatments that we consider to be promising. I think we should not do that to ourselves, either.

PARTICIPANT: I am heartened by hearing Meenhard tell us to stay with the course in terms of immunology. I am reminded 20 years ago by Louis Thomas that related to the cost, the encumbrances, the burden of halfway technology.

I am really brought back to think that, although CTLA4 adoptive transfer therapies are perhaps mechanisms to understand things, they are still a little bit like LAK, and we have all maybe done too much of that.

I don't think it is time to drop the cause for the specific interventions. I think we have just scratched the surface of this iceberg in terms of the T cell targets that have been evaluated.

Only the first three of the melanization markers have really been looked at, and we have a whole cornucopia of the rest of the CT antigens. We have all of the CD4 epitopes to look at.

I think that, whatever we do with the more cumbersome technologies, we really should still try to focus upon the simpler and, likely, in the end, higher technologies.

PARTICIPANT: I don't disagree with what Mike says, that by applying adoptive immunotherapy we get a lot of information, but I believe that will not solve the major problem that I believe exists for immunotherapy, and that is the development of escape mechanisms which continue to be utilized by the tumor cells.

It doesn't matter what type of immunity you use, whether it is adoptive, whether it is NK, whether it is antibody.

The major issue will still remain whether the tumor cells are or are not recognized by the immune system, and are or are not sensitive to immune distraction.

There is plenty of evidence today, I would say, starting from the immuno-editing information which was reviewed by Paul Chapman the first day, that the immune system can recognize the tumor cells, but then it is just a matter of time, because of the genetic instability and mutation and immunoselection.

I don't see that the problem is how to generate the immune response. It is more to understand how we can counteract the escape mechanism.

I would like to, in that regard, base our development of our strategy on advanced disease. In my opinion, that is major danger. The major danger is that the more the tumor cells exists in the patient, the more escape mechanisms they will develop.

In fact, if you look at loss of HLA, you will find that in melanoma, 60 percent of metastases have defects. If you go to primary, only 20 percent of the primary have defects.

So, by using the advanced patients, we are just going to draw conclusions that may be misleading, or may not help us identify the best strategy.

DR. URBA: We will hear from another non-immunologist. I am heartened also by the fact that the non-immunologist thinks we ought to continue because, if you sit in a room with immunologists all day like I did yesterday, we can think of lots of reasons why it didn't work and how it might work if we just changed something.

At some point, someone is going to have to look critically. I agree with Jeff, I don't think that we are there, where we ought to stop, but we can't test every antigen that comes up either, John, like you want. Let's hear from another non-immunologist.

PARTICIPANT: I would like to address the issue of cost. I mean, I agree that one shouldn't concern oneself too much with these in an academic setting, but it depends what the cost is.

If the major investment in the United States in melanoma research is in immunology and it is ignoring what else is going on, then I think you need to address that.

I think that the issue is, a year ago, there weren't many other markers or targets.

There are now, and I think we need to look at those.

I think that if you are going to continue to pour money into immunology and not study these other cellular pathways, then you are going to lose out. I am not in your cost pool and it doesn't really matter to me, but it does matter to me academically that this work is done.

KIM MARGOLIN: I would like to echo that. I was just going to mention that one of the things I heard yesterday from people who moved from group to group or switched when we were supposed to later on, was that there was a lot of overlap and possibly redundancy, particularly in the two target groups.

I think what I have seen, having spent my time in the high risk group, which we will hear from next, as well as your group at the end of the day, and what I am seeing here is that there is a pretty big disconnect between people who are thinking about immunologic targets and people who are thinking about all the other targets.

We oncologists really know that you can't expect any one pathway or any one approach to solve the problem. I think looking at the problem of escape from immune therapy, resistance to targeted therapy and heterogeneity of these targets, whether you look at the primaries, whether you look at the mets, we need all of these approaches to be combined, if we are going to get to this. So, we have all got to think together, I think, rather than thinking apart.

PARTICIPANT: I would like to present a somewhat different perspective of what happened in the last 10 years in immunotherapy, at least antigen specific.
I heard that there were some negative comments about antigen specific immunization, but if you really want to apply some kind of rigorous logic, there is nothing that has been as successful in melanoma than antigen-specific immunization.

Peptides, what they are supposed to do is just increase the presence of CD8 positive T cells that recognize tumor, and that is what immunization is supposed to do, and that has been incredibly successful.

I think the question is why these T cells can work. That is a different thing, and that is what we should focus on in the future.

Really, I think it is pretty naive to expect that these T cells that are exposed to an antigen every three weeks are going to work. That is not how it works in any viral model that I know of.

Acute viral infections that last for seven days, the T cell response then disappears within the next two or three weeks.

For example, the T cells we are studying are going through the memory phase. So, I think the future should focus on -- I don't think antigens really matter as long as they are expressed in the tumor cells, because if you believe in immunology, as long as you have a target, you have a T cell receptor recognized, and who cares what the antigen is.

I think that mostly we are understanding the physiology of T cells. I have to say that the viral community, HIV community, CMV community are much more advanced than we are in understanding the basics of T cell immunology.

The tragedy the last 10 years is that we have done very little to understand the biology of this phenomenon, particularly with looking at the effector phase.

The immunization, what the immunization is supposed to do is take care of the afferent loop, to bring T cells to the circulation.

Then, for the T cells to go there, having been inactivated and proliferated and work in the tumor microenvironment is a much more complicated thing that is unprecedented.

Really, viral systems are clearly different. The virus is not only the immunogen but is also the target. So, there is something that continues the immune process within the target organ.

I think that is a very important thing, that we should continue to immunize, but add something that is going to make it work, too, at the effector level.

MIKE ATKINS: I just want to strongly disagree with my classmate from high school, Dr. Sosman, and I can be as big a cynic about immunotherapy as you are, but I think what we have done is accomplished a lot, but we have redefined the questions.

We thought the questions were something different a few years ago, but we have come up with ways, as Franco was saying, of testing those, but what we really need is to design new experiments and clinical trials that address what we have now identified are the problems, and to use the tools that we now have that weren't available 10 years ago when we were doing these other experiments, to actually figure out what was happening.

Now is not the time to get disappointed and abandoning or favoring the next hot pathway that comes along, or molecule.

We know the road to effective treatment of melanoma is strewn with hot ideas that didn't lead anywhere. I think we need to keep pursuing this with the new technology that is available.

PARTICIPANT: It is interesting, we are talking about developing new targets.

Obviously, all the work that is being done with molecular targets in signalling pathways is exciting.

It is interesting, though, with melanoma and other cancers, we have targets. We have molecular targets that have been defined.

We have animal models that show that if you target immune responses to those antigens, you can eradicate tumors and/or prevent tumors.

Work still needs to be done in animal models, especially in understanding how best to immunize and those kinds of factors and studies on tolerance, et cetera.

We need to continue to push this in humans, and that is really where the questions are. What we realize increasingly is that the problem in humans is not to generate T cell responses, although we need to do it better, and it may or may not be to maintain the T cell responses, but to figure out how to deal with tumor induced immunotolerance and immune escape.

The people who have cancer developed cancer because they have immune tolerance to their cancer, and that is what we need to learn how to do, and we haven't yet done that, and that is an exciting area for the future.

Again, there are so many areas in medicine -- liver transplant, many other areas where people with normal levels of energy would have given it up long ago, but because some people are pushing it hard against the odds and against rebuke, it led to great success, and that is where we need to be.

PARTICIPANT: Let me say it again. I think the melanoma field is the leader in tumor immunology overall. There has been such a body of knowledge accumulated over the last 20 years, it would be at this moment, I think, rather foolish to say we don't need this any more.

The question is, one has to step back and follow a little bit the arguments that Franco was just giving, that one has to look at some of the mechanisms, what does or does not work.

Then there should be several ways of trying to combine the work on new targets, the ones Frank presented to us, with an immunological twist to it.

Then one can work up both the targeting as well as the immunology side of it. I think these are very fruitful investigations for the future.

DR. URBA: I think the committee would agree entirely with what you just said, and this slide was, I hope, to present what those mechanisms might be.

PARTICIPANT: I just wanted to make clear that I am not advocating jumping onto the next signalling pathway. These pathways are important in other cancers. So, they will eventually be studied.

So, I am not advocating dumping all the immunology in favor of going off and looking for BRAF inhibitors. All I am saying is that there needs to be a balance and, really, what you want to do is continue to spend the amount of money you are spending on immunology, but now try to get some more money out of the NCI to look at these other pathways.

DR. NISHIMURA: First, I want to point out that this is exactly what we dealt with yesterday all day in our workshop, this kind of discussion, which I think is very, very good.

I think all the points are right. I think where we are going to have to head with immunotherapy is very much the approach with other things, and that is combination therapies.

We do have to worry about targeting multiple antigens. We do have to worry about agents that make our T cells better, that can take into account concerns that Soldano was arguing about, and which Agusto and others will argue about.

I think it will not be a single agent that is going to be the cure for melanoma. I think it is going to be a multi-modality approach.

PARTICIPANT: I find this discussion very interesting. I will just point out to you two observations that got me started on this road 43 years ago, that are as valid today as they were then.

That is the fact that, melanoma, you do have complete regression for disseminated disease in one of about every 2,000 patients. The most likely explanation for that is the immune system.

That is a rare observation. A much more frequent observation is the recurrence of disease 20, 30 and I have even seen 40 years after treatment of the primary.

That can only mean that there is some mechanism in the patient that is keeping that tumor in check. The most likely explanation is the immune system.

Often, these recurrences are triggered by a major immunosuppressive episode. It may be a coronary bypass with a lot of blood transfusions that is immunosuppressive. It may be a major surgical operation, a prolonged illness.

It may be a death in the family and bereavement, which is very immunosuppressive. These are clinical observations that you have all made that treat melanoma time and time again.

That suggests that there is something that the immune system can do to control melanoma for long periods of time.

You know, this is about the fourth peak over the years of enthusiasm. Immunotherapy and immunology goes through a series of peaks and valleys.

You know, there was a peak in the early 1970s and then disillusionment, and then in the 1980s and so forth. I think we are going through a period now where everybody rushes in, does a few fragmentary trials, small trials and not well designed.

When you have got a tumor, melanoma, the average doubling time is 30 to 40 days. So, you start with a tumor that is two centimeters in size, that is eight billion cells that are doubling.

To think that you are going to get a patient immunized, which often takes two or three months, and cause that tumor to regress, is like pissing in the ocean. It is just not going to happen. It is unrealistic.

On the other hand, monoclonal antibodies, I first put monoclonal antibodies into patients in 1980. It had a great simple concept, a great idea, but it took 20 years before it really worked, because it is much more complicated than we initially thought.

I would just say hang in there. We are going to win. Immunotherapy is going to work in melanoma and probably other tumors. For you young people, it may take longer than you would like.

DR. SAXMAN: I just have what is probably a very simplistic question that I probably should have asked earlier, but it relates as well to your recommendation number one in terms of sorting out who is responding, the subgroup of patients that does seem to be responding to these various therapies.

One of the things that we heard on Monday that I was really struck by was how the genetic alterations in melanoma seem to be quite different based upon the phenotype of the primary.

I haven't heard anything further about that since then, but are we missing out on what would be a very easy opportunity there perhaps to sort out some of these issues as to who responds and what the genotypes are that respond, by not capturing that information about the phenotype of the primary?

I mean, John and Mike can correct me if I am wrong but, to my knowledge, none of the larger studies have captured that type of information and correlated that. Have they? Am I incorrect?

That would be something that -- Frank and Walt, do you think that would be valuable? That would be something actually very easy to do, it seems to me, and might be helpful in sorting some of these things out.

I just thought that was very interesting. I couldn't remember whether it was Richard or David that presented that on Monday, but that was information that I was not aware of, quite frankly, and may be valuable.

DR. HALUSKA: I agree with that, and someone said earlier that it is metastases that kill people, but the data are becoming clear that there is genetic heterogeneity between the different subtypes of melanoma.
It is also becoming clear that there is genetic heterogeneity that is correlated strongly with response or not to one or another therapy, not just cytotoxic therapy or immunotherapy.

The question is, do those things correlate with one another and the suspicion is yes. I mean, look at the history of a variety of other malignancies from lymphomas to, they all used to be reticulum cell sarcomas to the point where we came to understand that translocation set them apart, both diagnostically and therapeutically, to, even histologically, even lung cancers.

I think those genetic understandings have to be made, and just understanding the metastases as a homogeneous group is probably going to miss that opportunity.

DR. SAXMAN: If looking at the phenotype would help us figure that out, perhaps, that would seem to me to be very easy to do.
I am not saying that is the perfect way to do it, but it is easy, it is cheap.

DR. URBA: Mike has looked at predictive factors. I don't recall reading this one.

MIKE: We have actually done this in renal cancer with IL-2, where we have a bigger primary tumor to look at, and we have been able to identify, just with simple looking at the pathology things that might be associated with doubling or tripling of a response rate, and selecting out groups, more importantly, of patients who will never respond to therapy, just based on a histology review.

It is a little harder to do that in melanoma because the specimens are all over the place, and you don't have as much tumor tissue.

The logical step after that is to actually take frozen tissue or tumor blocks and try to identify features that are more molecular in those tumors that might predict for response to immunotherapy.

That is difficult in melanoma because a lot of where the primary gets removed is in a dermatologist's office, who is not going to go and snap freeze it, even if we had some way of actually studying it. He is not necessarily going to process it in that way.

I think we have not explored that. It would be potentially useful to identify all responders to certain therapies or long-term survivors and patients who were really high risk on an adjuvant study and try to get their pathology specimens to a few pathologists and review them blindly and see if there are any features that are clues.

We might want to look at where they are located, because it is possible that people with acral lentiginous melanomas never respond to a particular therapy, or someone with a lentigo maligna never responds, and we should never be treating those, and we should be treating those patients with some other approach.

It is still possible that our therapies actually work, but we are just treating the wrong patients. So, we can't see them. Maybe if we could identify what the appropriate treatments are for the right patients, all of our treatments would look better.

DR. URBA: A comment from pathology?

PARTICIPANT: I just want to remind you that that was part of the recommendations of group B, actually, to work on our understanding what actually genetically defines melanoma, and how many different types of melanomas there are.
I think that is something that be even done independently of what genetically determines a responder and a non-responder, especially if you really look at the true responders.

For IL-2, as you said, there are 10. So, these comparisons will be difficult to make. So, I think focusing on the primary tumors and see what is the heterogeneity of the disease, what are the genotype phenotype correlations and how many enemies are we tackling.

I completely agree with the analogy of lymphoma at the time where we had small round cell lymphomas. The clinical trials were not meaningful, and we need to understand what translations are associated, and that will also help us to identify the targets.

DR. WEBER: Walt, I think Mike made a very good point about response prediction. It is really an under-explored area, even with respect to chemo, chemoimmuno or any biologic therapy.

By analogy to other histologic subtypes like breast or colorectal, investigators are looking at DNA polymorphisms that might affect expression of important immunologic genes, IL-10, TGF beta, CTLA-4, and then DNA repair genes, apoptosis related genes.

That is an opportunity where you would just need blood to do a genetic analyses. You wouldn't necessarily need tumors, and that is under-explored in melanoma, and I think that gets to what Mike was talking about.

DR. EGGERMONT: I have two comments to make to the remarks by Mike. Of course, in and by itself it is attractive that this will lead to progress, and most likely, it will do so.

At the same time, the observation that biochemotherapy utilizing, say, up to six drugs is not providing survival benefit as compared to using DTIC alone or any of the other agents alone. That is a very worrisome observation.

If you could separate out effectiveness of various agents and they each, in and by themselves, would be successful in addressing a subpopulation, you would actually have expected that a six drug regimen would have provided survival benefit over a one-drug regimen.

So, there is really something very, very fundamentally wrong about our understanding, because the assumption that we are actually relatively successful in various groups, which then gets snowed under in the overall group, is a bit contradicted by the fact that six drug regimens are not successful either. So, there is something totally fundamentally wrong with our perception.

DR. MARINCOLA: Actually, I want to make a comment about the genetic analysis of melanoma. I am not so sure that it is going to work out as a predictor of response, because that is the assumption that you really have two diseases, and one is responding and the other one is not, based on a large number of genes that differentially express.

I believe that our experience is very limited. Unfortunately, we couldn't do a larger study than I hoped we could do.

We did fine needle aspirates before therapy and then we followed the lesions responding and not responding.

We found that, although there were different types of melanoma that were classed differently, none of them were predictive of response.

The only thing that seemed predictive of response were a very few genes associated with immune function, like the tumors are producing some cytokines or growth factors that seem to cause an inflammatory response.

I am not saying that is the answer but I wouldn't put all my eggs in that basket, look at other parameters as well.

DR. HALUSKA: I wanted to applaud Jeff for, I think, couching this discussion in the terms that he initially did.

It is interesting, I wear both hats. I have studied genetics and do clinical immunology, and have been frustrated with both of the endeavors.

We seem to not acknowledge that it is a zero sum game. If the majority of our emphasis is on one thing, it does exclude the other. I think we do need some balance.

It is quite clear that our field is reputed as one that the emphasis is on immunotherapy. On the one hand, that sets us apart as potential leaders in the field.

On the other hand, it distances us, in some respect, from one of the major thrusts of modern oncology biology, which is the understanding of the genetic changes.

The leaders in our field are not really studying that now. We are studying immunologic approaches, and I think we have to acknowledge that.

I think there is an opportunity now to readjust that balance a little bit, and that is not to detract from immunotherapy's promise or its successes.

We all acknowledge that some patients are responsive to very innovative approaches that we have come up with, but in both of the approaches, we have been less than successful, and I think this meeting has been a great opportunity to re-examine some of the assumptions we have made on both sides, and maybe to redirect our thrust.

PARTICIPANT: I think that is really important, going back to the points that Jeff made earlier about trying to get large pharmaceutical companies excited about this, and I think we have to deal with that.

PARTICIPANT: I would like to really just comment on Alex Eggermont's observation that multiple agents have not really been any better than a single agent.
I am a little bit surprised that the discussion hasn't centered more on the mechanism of killing of tumor cells.

An obvious explanation, I think, that comes to mind is that they are all killing by the same cytotoxic pathway. They are all inducing apoptosis through the mitochondria.

So, the chemotherapy is killing through the mitochondrial pathway, the immune system is killing through the mitochondrial pathway. So, when you combine all these agents, you don't get any additive effect, because they are all working on the same pathway.

I would argue that the way forward is to understand the mechanisms of cell death in melanoma cells a little bit better, look at the apoptotic pathways, why are they resistant to the apoptotic pathways, and target those resistance mechanisms.

DR. URBA: That is something that could come out of the DNA microarray.

PARTICIPANT: No, not necessarily, because a lot of these resistance mechanisms are changes in existing proteins, just phosphorylation of proteins. You will miss it in DNA arrays.

DR. HALUSKA: Why don't we take a 15-minute break. Then we will come back for the last session and wrap up the meeting.

TOP