Good afternoon. I would like to thank the organizers of this conference, in particular, Vern and Frank, for allowing me to speak this afternoon regarding the serum markers of relapse risk in melanoma. Hopefully, I can give you some details of what is known about serum and potential other markers that we can use to determine potential risk for recurrences subsequent, and also in response, to therapy.

TOP

In 1990, the concept of sentinel lymphadenectomy was introduced by Dr. Morton based upon the concept that we could identify the regional lymph node in direct connection with the primary tumor.

The technique was developed as a result of the controversy over elective node dissection, in order for us to identify, as surgeons, the single node that may be in the pathway, and identify those patients who have lymph node metastases, and also potentially prevent those patients who don't have regional lymph node metastases to be subjected to potential morbidity and cost of elective node dissection.

TOP

So, we have been able to identify the sentinel node and, through the use of routine H&E immunohistochemistry techniques, we have now been able to identify very small tumor volumes in the sentinel nodes, as shown here.

TOP

What has been shown by our group and others is that, although the sentinel node is very predictive, as you can see, in those patients that are tumor negative, the five-year survival from our group is estimated at around 90 percent, i.e., 10 percent of those patients who will have tumor negative sentinel nodes will subsequently die of their disease.

On the other hand, 70 percent of the patients with tumor positive nodes will remain free of disease and, therefore, the status of the sentinel node is not completely predictive from one patient to the next of their overall survival.

TOP

We have shown in multivariate analysis, as you can see, there are a number of factors where well-known Clark level, thickness, ulceration and including the presence of a tumor positive sentinel node as factors that we use to help predict the outcome of our patients.

As you can see, when a patient has a tumor positive node, we are still left with the other factors that may contribute to their outcome.

TOP

In fact, when we look at our patients who have tumor positive dissections alone, as you can see, some of the factors drop out, but the presence of ulceration of the primary, and certainly the tumor burden measured by the number of tumor positive nodes increasing, decreases the overall survival.

TOP

Yet, as surgeons, we are stuck with trying to identify, as Doug mentioned, about nine percent of the patients who may have disease in subsequent nodes.

Here is a study we did looking to try to identify those factors that may be predictive of a patient having more than one tumor positive node.

As you can see, none of these are absolute predictors of those patients. Again, as surgeons, we are left trying to contemplate which patient should have a complete node dissection versus those patients with a single tumor positive node that can be left alone after a sentinel lymphadenectomy.

TOP

On the other hand, as you can see, about 18 percent of our patients who have a tumor negative dissection will ultimately recur.

Our problem is that certainly, as clinicians, we can identify in transit and lymph node recurrences quite easily. Certainly the addition of ultrasound might improve this.

Our problem is trying to identify these patients early on, this 11 percent of patients who have distant metastases, and certainly modern radiographic techniques -- chest x-ray is probably the only thing that may be of benefit.

Certainly, PET and CT probably add little to our ability to detect patients early enough in their course of developing distant metastases so that we can have some change in their outcome.

TOP

So, the question is, where do we look to try to identify where we can determine, on an individual basis, what is going to happen with our patients.

There has been talk already this morning about the primaries. We have touched a little about the nodes, and also, my job is to talk a little bit about the serum.

TOP

Here are a number of markers that have been identified. There are certainly more. LDH is made from tumors, is also very non-specific, can be sourced from other tissues.

Melanoma inhibiting activity, we will go into S-100B, 5-(S-cysteinyl)-L-dopa is a product of the tyrosine activity through the melanin synthesis pathway, ICAM, etc., there are a bunch of them. We will talk about TA90.

Then, on the other hand, there is looking at immune related. That is, can we identify particular immune responses, measuring antibody responses as cellular T cell response, and certainly cytokine profiles, that may help us identify our patients, not only for risk of recurrence but response to therapies.

TOP

So, here is some of the data on LDH as a tumor marker that is measured by what is called an ultraviolet test.

As you can see, in a group of patients, these are stage III and IV patients. Patients had bloods drawn at very different times.
As you can see, with higher LDH levels, their survival was worse, suggesting from these authors that LDH may be a reasonable marker for determining subsequent recurrence.

The problem with all this literature is that it is a variety of different patients, having the bloods drawn at different times.

TOP

Melanoma inhibiting activity (MIA), originally described from a melanoma cell line, it is a 107-amino acid, 10 kilodalton protein. It originally was described for its ability to prevent melanoma cells from incorporating thymidine, but has now been shown to be related, that is, increasing serum levels. MIA has been associated with worse prognosis.

In this study, the problem is, as you can see, although this is probably reasonably acceptable -- 2.7 percent of false positives in healthy patients -- this protein also comes from cartilage, so this may be somewhat reflective.

As you can see, it is very difficult to determine certainly stage III and IV, if everybody expresses it, and only a minority of patients with stage I and II disease have elevated MIA levels.

TOP

S100B has also been looked at as part of the S100 family. In this study, again, an assortment of different patients, the problem being that, as you can see, when they did their multivariate model the log of the S100 failed in the model, suggesting that S100 in the serum, by itself, did not add anything to known factors that we already use for determining patient prognosis.

TOP

The TA90 is a 90 kilodalton glycoprotein identified in our laboratory at John Wayne. This was a group of patients of ours that were TA90 immunocomplex negative and positive.

As you can see, five year survival beyond is much better in the TA90 negative patients. Then, in this group of patients, by multivariate analysis, as you can see, the presence of TA90 immune complex and increasing thickness of the patient's primaries predicted overall survival.

TOP

As we move on to the question of immune responses, there have been a number of articles looking at cytokine responses in general.

The problem is that it is very difficult to get from the literature exactly what those responses might signify.

In particular, there may be a variety of different responses.

Some of these cytokines have very short half lives. They can be influenced by other factors such as cold or other illnesses.

This may be a better way. This is a group of our patients that received Canvaxin vaccine, but measuring delay type hypersensitivity, not only determining prognosis, but also response to therapy, may be another way that gives us a better handle on determining outcome of our patients.

TOP

We have looked by protein array, looking at it for a number of candidate different cytokines in our patients.

As described earlier, this is quite a mathematically complex matter to look at different cytokines, the different time points.

TOP

What we did, we looked at a patient group that was very well matched. These were stage four patients matched for known prognostic factors.

We separated them out into the highest quartile survival versus the lowest quartile. As you can see, in our patients we found a much higher IL13 level in those patients who had longer survival, and then those patients with poor survival had PDGF beta found in their serum, suggesting that these may be candidates to look at in the larger prospective databases to determine the outcome in the patients.

TOP

As I mentioned, the problem with looking at cytokines in the serum is there can be a number of different cytokines, some increasing, some decreasing, it depends what kind of immune response is occurring at that moment in time.

They can be influenced by a number of different factors and we don't know really which one we want at which particular time.

So, the limitations of measuring serum cytokine is certainly questioning exactly what it means at the time you are measuring it.

TOP

Now we move back to the sentinel node, where we think that there may be more information as really the microenvironment of the first tumor metastases.

We have moved from the era of pathologists looking in the microscope to now the molecular oncologist.

TOP

Here is data that was just presented by Dr. Morton at the American Surgical, demonstrating similar outcomes to what Dr. Rankin has illustrated, in those patients who underwent exhaustive analysis of their sentinel nodes. They were H&E negative.

Also, these were measured by PCR of the paraffin section. So, this really allows the next step. It no longer has the pathologist competing with the molecular oncologist to try to identify the tumor metastases.

Now the pathologist gets their fair share. Now the molecular oncologist gets their fair share. As you can see, in follow up of greater than six years, the overall survival is around 95 percent, and then adding the presence of one of these four markers, not only melanoma associated markers, but also related to cell cycle regulation. We have now found that that is a strong predictor of outcome in these patients.

TOP

Certainly, there has been work just recently published suggesting similar ideas, that survivin, which is a regulator of cell cycle regulation apoptosis may be important.

In this Italian study, as you can see, in those patients who did not express survivin in their sentinel node, that 100 percent of them survived. Although the study is small, it certainly has had reasonable follow up, suggesting that maybe not just melanoma associated markers, but markers of cell cycle inhibition, anti-apoptosis, may be good markers to look at for regulation and what is happening in the sentinel node.

TOP

We have focused more on gene expression of cytokines and we demonstrated a year ago that we could identify cytokines by the use of semi-quantitative PCR in sentinel and non-sentinel nodes.

 

TOP

Now we have gone on to look, by quantitative PCR. This is a case where the primary melanoma was intact, sentinel node negative.

As you can see, the cytokine expression from the sentinel node where the primary is intact, and in the next case, this is where the primary is gone but now there is metastases in the sentinel node, a different patient, but different profile of cytokine expression, again, demonstrated by quantitative PCR.

So, this is at the gene level. We obviously don't know exactly what the protein level is, but it suggests the microenvironment of the tumor may be important for us determining ultimate outcome of our patients.

TOP

Certainly, there are other places to look. There have been a number of studies that have already looked at the acellular portion of blood, looking at melanoma specific gene products, in particular, tyrosinase, gp100, etc..

These studies have become very controversial. There is a lot of question about the techniques different people used, the timing of blood draws, et cetera, which have made this somewhat controversial.

Certainly, an interesting way to look at it is to look at telomerase activity. Again, this is something for the future. Certainly, looking at the cell portion of the blood, looking at oncogenes, tumor suppressor genes, etc..

We have a number of different methods of looking and certainly we have a number of different pathways and compartments to look at.

TOP

Overall, we think the future is bright with the sentinel node, not only for determining certainly patient prognosis alone by H&E, but also by newer molecular oncology techniques.

We talked a little bit about the serum markers that have been used and, frankly, there have been none that have panned out completely. None have been validated in any large clinical trials.

Certainly, development of new methodology in order to identify potential candidate proteins and genes may be worthwhile.

Certainly, we still believe the sentinel node technique has a future, and certainly a big part of our clinical practice and understanding of the pathway of metastases of melanoma. Thank you for your attention.

TOP