Melanoma 2003

SLIDES & TRANSCRIPTS
Monday, May 5, 2003

Commentary: Con

Alexander Eggermont, M.D., Ph.D.

Slide 1:

When I received the eight pages of instruction by Vern, I decided to respond in the same way that Kelly McMasters did, and go into a full denial that I ever received the eight pages.

So, all I did, before going on a trip abroad, was e-mail to Vern telling him that I noticed he had put me under contract as a discussant, and that this was taken by me as an instruction to deny everything that had been said previously in the session, and that is what I will try to do.

Of course, I did not open my mail box since coming back to see what Vern had cooked up.

Utility of staging prognostic factors for melanoma.

Slide 2:

In other words, is the glass half empty or half full? This goes by somebody whose glass is usually half empty.

Slide 3:

Now, we know that primary tumor has a number of important factors -- Breslow, ulceration, mitotic index, vertical growth factors in thin melanomas, lymphocytic infiltrate, especially the Morrow and Meemuth(?) studies.

Slide 4:

So, we know all this. We know the importance of Breslow and Clark.

Slide 5:

Breslow being the more important one, and we know the importance of ulceration. That has all been discussed.

Slide 6:

The interesting thing is that, out of this melanoma unit, there will be a publication coming out in Cancer on thousands of cases where they actually scored for mitotic index rather than ulceration, and where, as I understood it in the multivariate analysis, mitotic index comes out as the important prognostic factor, rather than ulceration.

Vertical growth factor in thin melanomas, of course, remains a hallmark of aggressiveness, and there will be a large study on lymphocytic infiltrates on the primary melanomas, where the UACC has joined with Mimna(?) with a study initiative and will hope to come in with well over 1,000 primary melanomas to be studied for a lymphocytic infiltrate. That is all primary melanomas coming out of our randomized phase III trials.

Slide 7:

Regional lymph node status, of course, is the next and very important prognostic factor, because there we have the first proof of a metastatic phenotype.

When we discuss microscopic involvement only of regional lymph nodes, we have elective lymph node dissection data and we have sentinel node data.

Slide 8:

Now, regarding elective lymph node dissection data, we know that all historic studies and all non-randomized studies tended to demonstrate a benefit for elective lymph node dissection. Of course, historic control studies always do that.

When you look at non-randomized studies that compare patients studied in the same period of time, not the historic controls, and compare elective lymph node dissection to non-elective lymph node dissection patients, already there are three large studies that show that the elective lymph node dissection was not having a big impact at all and the randomized trials, by and large, were negative.

Slide 9:

Craig Slingluff published, when he was still at Duke University, a database which showed that there is a remarkably similar percentage of patients who have distant metastases and patients who have lymph nodal metastases, regardless of the thickness that they come from.

Slide 10:

The Sydney melanoma unit, as a matter of fact, did a long-term follow up for patients who did receive an elective lymph node dissection or did not receive elective lymph node dissection over one period of time in their clinic. They failed to show any evidence that elective lymph node dissection improved outcome.

Slide 11:

Basically, the intergroup trial on elective lymph node dissections in the United States failed to show an important effect of doing elective lymph node dissections. So, basically, the failure of adjuvant surgical procedures to improve outcome.

Slide 12:

An important analysis was done by the first WHO trial on elective lymph node dissection, and actually we now understand these data much better than before, based on what we have learned from sentinel node stage.

The interesting analysis here with very long follow up is that they looked at the pathologic staging outcome of the elective lymph node dissected patients.

For that, you need to look at these two curves. The green curve is the ones who underwent an elective lymph node dissection and were found to be nodal positive, microscopically.

They basically had the same survival curve as the ones who were not electively lymph node dissected, if that is correct English, but who were undergoing delayed lymph nodal dissection because of palpable nodes.

So, these two curves are identical, and therefore, would show that it didn't make any sense to do elective lymph node dissections because you could not improve outcome by doing the lymph nodal dissection earlier. So, these are two very important curves.

The two upper curves are the ones that never underwent an elective lymph node dissection, and these are the ones that underwent an elective lymph node dissection, but were found to be node negative.

The interesting thing is that this difference actually was significant, and it led to the hypothesis that patients actually had a disadvantage by undergoing elective lymph node dissection and all sorts of immunological mechanisms were hypothesized to explain this.

Of course, what we know now is that this is probably a reflection of false negatives in this negative lymph nodal population, because these were all staged the old style. If they had been staged more appropriately, many of these would have actually gotten down to the stage III stage.

This, of course, would not happen as frequently with sentinel node staging and the work up of sentinel nodes.

Slide 13:

Now we are in the era of sentinel node staging, which has had enormous implications in the management of breast cancer, and there has avoided the full lymph nodal dissection in about two thirds of the patient population, as compared to what we used to do five to 10 years ago.

Slide 14:

Sentinel node staging really re-defines stage two patients in melanoma. You are either sentinel node negative and you have a very good prognosis, or you are sentinel node positive, and you have a relatively poor, or you have quite poor, prognosis. So, stage two, in a certain sense, does not exist any more.

Slide 15:

Now the question, then, about sentinel node staging is, does it have a direct impact on outcome, or is it just a tool for stratification.

Of course, what I need to reflect on is whether any of these ultrastaging methodologies is changing the outcome of patient treatments, rather than changing the way that we are looking at biological phenomenon, and whether sentinel node staging has a direct impact on outcome.

It is not likely, when you look on the elective lymph node data in melanoma, on the previous phase III trials, and also in breast cancer, the first trial which was done, which randomized between doing yes or no. Full axillary node dissection was negative for that intervention.

So, there is a lot of data around to show that it is not very likely that you are going to significantly improve outcome by doing a regional lymph node dissection, but of course, the true test here is the sentinel node trial by Don Morton, and we hope to see an analysis of that trial within one year, I assume.

Is it only a tool for stratification, stratification for trials, to get more homogeneous patient populations to be more effective in analysis of treatment outcome and could it, thereby, have an indirect impact on outcome?

That, of course, then assumes that we have effective treatments. Of course, we all know that, in melanoma, this is exactly what we are lacking.

We are lacking effective systemic treatments and there is not basically a treatment around with solid evidence that it could salvage more than five percent of the patient population under study.

Slide 16:

Regarding sentinel node staging, it is extremely important in terms of how you work up the sentinel node. This happens to be a patient of mine who had non-palpable nodes. He had a primary with a vertical growth phase, not ulcerated, 1.5 millimeters on the back.

She had two nodes here on the scintigraphy in the right axilla and one in the left axilla. The sentinel nodes in both axilla were positive, but only on immunohistochemistry.

So, I did a full lymph nodal clearance on both axilla. There, I asked the pathologist to work up the highest nodes in both axilla as if they were sentinel nodes, but I kept them separate, and I kept them after he had done the full analysis of some 30 nodes that came out of each axilla. All the other nodes were declared negative for tumor.

Then I asked him to work up the highest nodes, which I had kept apart, kept separately, and work them up as sentinel nodes. Both the highest nodes were found to be positive, when worked up as a sentinel node.

Basically, you would have here patients with very high likelihood of systemic disease throughout the body, even though the initial outcome of your pathology would have suggested perhaps a mitigated version of melanoma and, unfortunately, this patient indeed followed the course of the prediction by the sentinel node work up of the highest nodes, and she died within one year.

Slide 17:

So, where in breast cancer -- this is another point that has not been discussed but I would like to bring this up -- where in breast cancer we take usual step sections through the node.

The melanoma sentinel node gives the best yield of work up pathologically by focusing on the central plain, bivalving the node through the highland.

Slide 18:

[No text is associated with this slide].

Slide 19:

Here, I would like to add some data that come out of the UHC melanoma program that will be published one of these months in the Journal of Pathology.

What we did, we bivalved the nodes and we focused on the central plain and we took step sections only out of the central plain by 50 um steps, up to a maximum of 700 um.

Slide 20:

[No text is associated with this slide].

Slide 21:

When you work on the melanoma central nodes that way -- and this is a study in 1,200 patients, and you work them up by bivalving them through the highland, then in the central plain you do your step sections of 50 um over this maximum range.

We actually increased the positivity by H&E and immunohistochemistry only from 17 to 34 percent, and there were no false positives by this evaluation.

Now, by adding RT PCR to this method, we actually scored another 10 percent of lymph node positivity. However, seven percent of those were false positive, because it was RT PCR on the frozen sections.

These seven percent were all nevi cells in the central nodes, although the RT PCR only increased the positivity by three percent.

So, this actually means that this is a crucial factor and that we do not necessarily have to go all the way with RT PCR of central node evolution because it is very difficult to get rid of seven to 10 percent false positives that way.

Slide 22:

[No text is associated with this slide].

Slide 23:

[No text is associated with this slide].

Slide 24:

[No text is associated with this slide].

Slide 25:

[No text is associated with this slide].

Slide 26:

These things have already been shown and therefore I will not comment on them.

The other major difference in the prognosis is, of course, whether you have microscopic disease only or you have macroscopic lymph nodal disease.

Slide 27:

The importance is, if you have microscopic disease only, this issue is metastases free survival, and if you have palpable nodes, this is significantly worse.

Slide 28:

[No text is associated with this slide].

Slide 29:

This has been shown to be true, then. This further differentiates out in high risk primaries thicker than four millimeters, with this metastases free survival curve.

This is for sentinel node positive patients, by and large, or one positive node in the palpable nodal patients, and these are by and large palpable nodal patients, two to four nodes, and more than five nodes.

Slide 30:

So, this is all extremely important prognostic data in terms of overall survival. All these curves separate out very nicely.

Slide 31:

What we did, in our 1995 trial, an intermediate dose interferon trial, was we looked at sentinel node patients, the red curve, one node positive or sentinel node patients multiple node positives, versus palpable nodes, one node, palpable nodes, two to four nodes, palpable nodes four or more nodes.

Thereby, we were showing that the sentinel node patient population later on basically joined the one positive palpable nodal patient population that is out there.

Slide 32:

[No text is associated with this slide].

Slide 33:

So, phase III by and large is still very heterogeneous patient population. The largest different is one versus N2 and further number of positive nodes.

The tumor load in the sentinel node patient population. I will show that later by looking at the starts classification, which has not been discussed yet.

Still, alteration of the primary or Breslow thickness still has some remaining impact,

Slide 34:

but the evaluation of the sentinel node by the staging system of starts from Alpsbourg(?) in Germany, which is like a Breslow thickness type of indication of the tumor involvement in the sentinel nodes, separates out, greatly again, survival curves.

So, you can identify patients who may have a few cells in their sentinel nodes, but who will not go on to metastasize. So, there is still a lot to be learned there.

Slide 35:

Especially if you then further combine it with tumor thickness of the primary, you have a very bad patient population and you have relatively good patient populations.

Slide 36:

[No text is associated with this slide].

Slide 37:

So, stratification practice now days for trials in stage III patient populations then need to look at microscopic versus macroscopic, the number of nodes and these stratification factors and, in stage II, the most important data is are they sentinel node stage, yes or no,

Slide 38:

because if they are node negative, they have a much, much better prognosis.

Slide 39:

[No text is associated with this slide].

Slide 40:

[No text is associated with this slide].

Slide 41:

So, the next item, then, is RT PCR for circulating tumor cells. In spite of the promising data shown, when you look at the about 200 studies published up to 2003, the situation has very little changed before the metaanalysis published in 2001 and after.

Vertow(?) et al from Harvard have published in the Archives of Dermatology a review on one of the 27 studies -- 50 were reviewed, and 23 were good enough to be actually analyzed.

They find that, in stage I, II, III, IV, you see an increasing percentage of circulating tumor cell positive patients, but you see that in stage IV it is still only 45 percent.
This makes it a very discontinuous type of event, and so, meaning that the applicability of RT PCR for circulating tumor cells is still very questionable.

Slide 42:

Now, another 73 studies have been published, but whereas quantitative PCR may have shown that it has important implications, how you treat your samples and the quality of your sampling may influence the results.

So far, there is no clear change in the applicability of RT PCR for circulating tumor cells compared to the meta analysis situation in 2001, and there still remains methodology on which you cannot really base clinical decision making, but it is a research tool. Of course, the outcome in the melanoma Sun Belt Trials will be important here.

Slide 43:

Regarding prognostic serum factors, there are antigen based prognostic serum factors, S100, TA90, MIA, but whether these are truly independent prognostic factors, the evidence for that is very thin.

Slide 44:

Prognostic serum factors, you have a whole bunch of other ones, LDH, melanin based, matrix based, inflammation based, cytokine receptors and adhesion molecules. There is no evidence around that these are truly independent prognostic factors.

Regarding prognostic serum factors, there are antigen based prognostic serum factors, S100, TA90, MIA, but whether these are truly independent prognostic factors, the evidence for that is very thin.

Prognostic serum factors, you have a whole bunch of other ones, LDH, melanin based, matrix based, inflammation based, cytokine receptors and adhesion molecules. There is no evidence around that these are truly independent prognostic factors.

Slide 45:

So, is ultra staging, then, meaningful in the absence of effective adjuvant surgery, marginal and elective lymph node dissection?

Is it meaningful in the absence of effective systemic therapy in stage II, III, IV?

If we would go along with the premise that we don't have agents that cure more than two to five percent of patients, be it DTIC, be it interferon or be it IL-2, or even the combination of that, there is no solid data that this combination will actually cure more than five percent of patients.

Slide 46:

So, does it change treatment? Is it meaningful for the patient? For stage two and if, after sentinel node staging, remain a stage II patient, it will only have ramifications in terms of trial participation.

If you are upgraded from a stage II to a stage III patient from central node staging, you may receiving interferon but, in Europe, you may not. You may be randomized between interferon or any other adjuvant therapy protocol versus observation.

It will, therefore, have implications for trial participation. It will provide you with cleaner trials.

For stage III IV, if you are upgraded from a stage III to a stage IV by your ultra staging, it will only mean that it may change the routing in trial participation, whether you receive interferon or vaccines.

Slide 47:

So, does it change outcomes? Sentinel node, does it change outcome in patients? It identifies early stage III. There may be fine tuning by the starts system, the SI, II, III system.

It is a stratifying tool for trials, yes, but there is no evidence of a capacity to influence outcome at this point of the game.
RT PCR, it is a research tool only for circulating tumor cells.

Serum tumor markers, it may identify high risk for relapse. It may be able to monitor response, but there is no evidence that either capacity influences outcome of treatment in patients.

Slide 48:

So, you tell me, is the glass half empty or half full? This really still remains on the table and I tend to say, in the absence of truly effective systemic treatments, the glass remains half empty, but it is half full in terms of better understanding of the disease and using this methodology in terms of finding out how to do cleaner trials. Thank you very much.