My title has also changed, because Dr. Sondak also asked me to convince you that we still need pathology in our molecular world. I am convinced that we do.

TOP

There are a couple of things. Some of you may recognize Dr. Ackerman and Dr. Mihm here, peering over the microscope together.

I think probably the most important thing at the outset is whether we are dealing with melanoma or not. You certainly need a pathologist to make that determination. In some cases, clinically, it is almost obvious, but I think we all feel the most comfortable in the setting of histopathologic examination of the tumor.

Really, what we are asking is whether these tumors have metastatic potential. There are controversies about whether Spitz tumors metastasize, but I think we are really talking about metastatic potential.

Even more important, at the time of staging, what is the likelihood that the tumor has spread beyond the primary tumor site.

Another way of putting this -- and this is what we are talking about today -- is how do we identify those patients with localized melanoma -- i.e., stage I and II disease -- that harbor minimal residual disease after local excision.

TOP

This is a benign dermal nevus. I think most of you will recognize that, and certainly those of you in the front will see that the superficial cells are epithelioid, they have pigmented cytoplasm and, as you move more deeply through the tumor, we have less cytoplasm in the cells and smaller nuclei. Some of these at the base even look like lymphocytes.

I think it is important to remember that you need to distinguish the lymphocytes from melanoma cells at this deep aspect, or melanocytes at this deep aspect, and when we get to some of the markers, you will see how important that distinction is.

TOP

Now, this is a nodular melanoma. Its scanning magnification, it might bear some resemblance to that dermal nevus but, with more careful examination, we will identify the characteristics that allow us to distinguish this as melanoma.

This is nodular melanoma. That was a benign dermal nevus. I have presented the most straightforward of possibilities in melanocytic tumor progression, but I think even here you can see that there is some overall resemblance.

The first patient has a scar to show for their tumor. This patient has since died of metastatic melanoma.

TOP

Primary cutaneous melanoma. Well, what is important? I think an important histologic prognostic factor might be something that has some statistical significance in determining prognosis.

All of these things that we include in our pathology report have been shown in studies to have a statistical significance with outcome, whether it is disease-free survival or overall survival.

Most of these, we know, are very tightly linked with tumor thickness. I think what we are asking today is, what are the factors that will affect care of the patient.

Well, certainly as we will hear from Dr. Buzaid, the tumor thickness, ulceration and Clark level are important in staging early stage melanoma.

What gets used in the care of the patient really, I think, is not uniform from institution to institution, or even patient to patient.

I think there are some of us that know that a patient with a very mitotically active tumor that is thin might get a sentinel lymph node dissection, whereas that same patient with that same tumor characteristics and no mitoses may not.

So, what we include in the pathology report is all this stuff. How it is used is not entirely uniform.

TOP

Just to mention, in the new AJCC staging schema, early stage melanoma is entirely determined by three histologic prognostic indicators: tumor thickness, which we now fortunately have very nice, easy cut offs to remember. They are millimeter segregations.

The presence or absence of ulceration, I know we are going to hear more about that. Then, for the thin tumors, it is important to note the anatomic level, and in a very invasive tumor that is less than a millimeter thick, this will be upstaged.

TOP

This is a chart showing you the importance of this kind of pulling together multiple histologic factors to segregate patients into subsets. This is, here, based on five-year survival.

What I have done in this chart is to compare with the prior AJCC schema the pool of five-year survival. You see that there certainly is some correlation, but in the new schema, we are able to break things out a little bit more.

TOP

Ulceration, I show this picture just to remember to note that tumor thickness is very important, and it is important to know how to measure tumor thickness.

In an ulcerated tumor, you don't measure any of this stuff. You have to find the top most viable melanoma cell, and then measure from that to the deepest melanoma cell.

TOP

Here is a schematic showing us this, from the top most viable cell to the deepest portion of the melanoma. Remember, the histopathologist will not measure that is going down around adnexal structures, nor do we measure tumor thickness based on microscopic satellites that are distinctively separated from the primary tumor mass.

 

TOP

Dr. Clark described the levels of invasion. This is just such a beautiful example of the difference between level III and level IV.

This is from a Cancer paper 30 years ago, and shows this infiltration through the reticular derma collagen and into the fat in level IV and V melanoma.

These are the tumor levels that patients are upstaged, if they have less than one millimeter thick tumors, compared to the more pushing border and less infiltrative, almost less invasive border.

Remember, anatomically, the distinction between this level III and level IV is a vascular web or plexus that separates this reticular dermis from the papillary dermis. So, this is a very anatomically significant finding observed in that H&E section.

TOP

Mitotic activity has also been shown to correlate strongly with patient outcome. Here, we have a rather dismal eight-year survival in patients that have tumors that have more than six mitotic figures in a millimeter squared.

TOP

Tumor infiltrating lymphocytes, we are going to hear a lot more about this in the upcoming years. I am not going to go into it in too much detail, other than to note that the definitions are a little bit cumbersome.

TOP

But, if you adhere to these definitions, it has been shown in two large and very well characterized studies that the nature of the host inflammatory response to the primary tumor does correlate.

The pattern and distribution of this infiltrate does correlate with outcome, whether you are looking at five-year survival in this study or the 10-year survival in the more recent WHO study.

TOP

Regression, another cumbersome histological feature but, nevertheless, does correlate with outcome in patients with primary cutaneous melanoma.

TOP

Then, just to note, the microscopic satellite is one that is present in the section with the thickest primary tumor.

Here is the primary melanoma. This is distinctively offset from the primary tumor, not contiguous with it, and this nodule or focus measures by definition more than .05 millimeters.

TOP

So, our gold standard remains tumor thickness. There are a number of other things. Some of them we now include in staging very regularly, but all of them are used in conjunction with primary tumor thickness.

TOP

So, what I would like to do in the next few minutes is present a paradigm for the use of a molecular finding in the determination of prognosis.

To date, there are no molecular tests that we use in our clinical approach to patients with localized melanoma.

All of our staging and treatment really is pretty much based on the measurement we make with the ruler, looking at stained tissue. That is pretty remarkable, given all the computer stuff that we have heard about this morning.

TOP

This is a description of the discovery and what we have found out about melastatin. This is a gene that was discovered by a differential display of mouse melanoma cell lines.

The high metastatic cell lines had a loss of this gene. Of course, many genes were looked at, but melastatin is one that we identified that was down-regulated.

This is work done with Millennium Pharmaceuticals in Cambridge many years ago. In our pilot study, we looked at in situ hybridization using S35 labeled riboprobe on a number of human melanomas. These were routinely processed tissues.

All of the benign melanocytic proliferations showed ubiquitous high level expression of melastatin message.

Many of the melanomas showed ubiquitous expression of melastatin message, but some of those primary melanomas showed focal loss, or loss of melastatin, and every one of our metastases showed loss of melastain expression. I thought that was pretty interesting.

TOP

This was a dermal nevus in our study strains with ubiquitous expression. This is a dark field. It is S35 labeled riboprobe, so you dip this in photo emulsion film and the white grains correspond to signal. This is our dermal neva showing diffuse expression of melastatin message in this benign nevus.

TOP

This was a dermal metastasis of melanoma showing complete absence of expression of melastatin.

TOP

When we looked at those primary tumors in the pilot study, we also saw that there was correlation of loss of melastatin message with increased tumor thickness.

In looking at these few patients, it was interesting to me that this patient, who had a very thin tumor was dead, unfortunately, within one year of primary tumor diagnosis, of metastatic melanoma.

This patient, with a 2.9 millimeter melanoma, who had diffuse expression of melastatin is still alive 10 years after diagnosis of the melanoma.

TOP

This is an example of a dramatic loss. Here we have primary melanoma. This is also melanoma and, over here in the dark field, you can see there is loss of message in this region, and really, blazing levels of messenger RNA here.

TOP

This is a little bit more subtle case. Again, most of this is melanoma. Some of you will note that there are some lymphocytes in here, a few lymphocytes here. If you look at the in situ hybridization, you will see very high levels of message expression, but there are some foci in here that you might question about whether there is tumor there and no message.

Sure enough, when you look more closely, again, here are some lymphocytes that are negative, but we do have tumor cells here that are not expressed in melastatin, and we do have tumor cells here that are expressed in melastatin.

Now, I have shown you one of our more difficult cases just to show you how important it is for you to have some skill in looking at slides when you read this test.

TOP

I won't go through all the details of this, it is published in the JCL, but if you look at melastatin lost and melastatin present, an easy segregation of these 150 patients in our study, you will see a correlation with outcome.

As I noted before, it is very important to do correlation with tumor thickness. These patients were staged by the prior AJCC scheme. So, we are really looking at patients with thin melanoma versus patients with thick melanoma.

TOP

Patients with loss of melastatin and no loss of melastatin.

If you have a thin tumor and your primary tumor ubiquitously expressed messenger RNA for melastatin, there is a uniform chance of eight-year survival in our study.

That is compared to a 50 percent chance of being alive if you have both a thick tumor and loss of melastatin.

This was sort of interesting. If you look at the patients with the thick tumors but no loss of melastatin, they had about a 90 percent eight-year disease-free survival rate as opposed to patients that had thin tumors that did have a focus of melastatin loss.

TOP

In summary, this is some material that I didn't present, but this is melastatin. It is a gene that is present on chromosome 15, and is believed to be a calcium channel regulatory protein.

We heard earlier today that Mitf may bind. This is still, a lot of work needs to be done to better understand the function of this protein.

We feel strongly that this is a marker of melanocytic tumor progression. It is present in benign nevi and melanoma in situ. It is lost with tumor progression, and the correlation with cell line with metastatic potential exists.

TOP

We also have correlation of the melastatin message expression without melanoma.

Loss correlates with an increased risk of metastatic disease and, in multivariate analysis, we found that melastatin message expression, tumor thickness and mitotic rate were independent predictors of prognosis in these patients.

TOP

I wanted to bring this up because it seemed relevant. This is shifting gears completely, but this is the melanoma protocol that we use at Massachusetts General for the evaluation of sentinel lymph nodes.

We do Mart1 and S100 and, in our study -- well, going back we found that this allowed us to detect 12 percent of the patients with negative lymph nodes on the first H&E, and 10 percent of our patients with sentinel nodes have nodal nevi, and we found those on the immunohistochemical studies as well.

TOP

[No text is associated with this slide].

TOP

[No text is associated with this slide].

TOP

Finally, just to summarize, these are the factors that we issue in our pathology report. We will keep doing that because we think it remains very important in the management of these patients.

Melastain, I presented, is a paradigm for molecular prognostic markers. Currently, I just wanted to put a plug in for the CALGB trial that you will see listed in your list of trials that are in your packet.

This is a study correlating melastatin with sentinel lymph node status. Thank you.

TOP