Melanoma 2003

SLIDES & TRANSCRIPTS
Monday, May 5, 2003

The Future of Staging and Its Implications for Clinical Trials in Melanoma

Antonio C. Buzaid, M.D.

Slide 1:

First of all, I would like to thank the organizing committee for the honor to be here among such a remarkable group of scientists and physician scientists.

My task, after receiving the eight-page Vernon's assignment, he had a number of subtle questions. One is, why are we here on earth. That was the most difficult one.

Now, my goal here is to briefly touch on the past, present you the current staging system and outline how I foresee the future for the staging system of melanoma.

Slide 2:

Now, a brief look at the past, I think, would really be considered very interesting. This is an old paper. This is 1968. Most of us were born at that time.

It was 650 patients. One of the investigators, a plastic surgeon, divided the tumors in a very simple way. Either it was good or it was bad.

Slide 3:

It was primarily a local problem and it was called, a typical example, lentigo maligna.

Slide 4:

Now, in a bad one, it was a systemic problem. Ulcerating melanoma was already there. This is 1968, and the satellite lesion is on the right, a big lesion on the left. This inspired a new staging system, and I will show you how very soon.

Slide 5:

Now, this evolved markedly over the years. In 1983, the staging system used primarily the cut offs reported by Dr. Breslow in 1970, 0.75, 0.76 to 1.5 and so forth.

Finally, based on a large data set, it was modified to a more simplistic way to effect a better fit in terms of survival distribution, and ulceration was added to the staging system.

Slide 6:

Now, the nodal category was also modified markedly along the years. In 1983 the cut off was five centimeters, and if it was greater than five, it was already called an M-1.

In 1988, based on nothing, it was modified to less than three or greater than three centimeters, and obviously, it was obvious.

The number of positive nodes were finally incorporated into the staging system. Whether it was microscopically evolved or macroscopically affected became also a modifier of the N category.

Slide 7:

The advanced disease category, we haven't done very much, in my opinion. It might be a little more confusing now, but we haven't added that much.
LDH, serum LDH, is now part of the new staging system, and it was before only for testes cancer, and now this is the second cancer in which a seromarker is part of the staging program.

Slide 8:

Now, I am going to show you another touch on the past before I go to the future. The most striking paper, in my opinion, ever published on prognosis of melanoma -- this is from Memorial, 1953. I wasn't born. I was born five years later.

I will take the liberty of reading this piece. The significance of ulceration. It has long been appreciated that ulceration in a mole is an ominous clinical indication of cancerous change. In a general sense, this impression is true, but several qualifications should be mentioned.

Slide 9:

Furthermore, on this paper, as might be further expected, the incidence of ulceration in the primary melanocarcinomas -- the way they were called in 1953 -- is greater among the fatal cases than the survivors.

Slide 10:

The final conclusion of this paper by Dr. Allen and Spitz, this is very striking. It is a precursor of Breslow.

It is to be concluded, therefore, that while it is impossible to predict which lesions will prove fatal, the odds are sizeable that the patient with the superficial melanocarcinomas, in locations other than the mucosa, has considerably more chance of survival than the patient with a more deeply invasive tumor.

Moreover, it was noted that it apparently takes only a small added increment of depth of invasion to cause sharp deterioration in the group prognosis. Dr. Breslow reported the classical paper in 1970.

Slide 11:

A big leap now to the future. Dr. Balch put together an incredible database that assembled European data, Australian data and U.S. databases.

This database set the new staging system. This is a 17,000 patient database. Most patients had primary disease, around 15,000 patients.

Of great interest, about 5,000 of them had the regional node status assessed histologically, either by regional node dissection or sentinel node biopsy.

Slide 12:

Patients with palpable node was around 600 or so, and then about 1,000 patients had metastatic disease.

Slide 13:

I will go over some of the most important factors of the new staging system, which comes from this data base.

Cox regression analysis shows that, for patients with primary melanoma, without considering the sentinel node or histologic status of the regional node, thickness and ulceration were the two most powerful factors.

Slide 14:

A number of other factors were just described by Dr. Duncan and many others, and some of them are very promising but, in my opinion, not yet ready for prime time.

The databases that they are based on are relatively small, and it should be validated in large data sets.

Slide 15:

One interesting finding that my colleagues at M.D. Anderson and I originally reported in 1997 was, when we asked the statistician, Sanjio Sung(?), what happens when the best fit was one, one to two, two to four and greater than four, and we asked him, what happens when you plot ulceration for each category.

To our surprise, it fell exactly into the next category. The ulcerated are the dotted lines, the solid lines are the non-ulcerated group.

Let's get an example of a patient with one to two millimeters, the blue line. If it is a dotted line, it falls exactly in the two to four millimeter group which is not ulcerated. This was reproduced in this 17,000 patient database.

Slide 16:

Of relevance in terms of, we want to make the system simple, and it wasn't that easy to try to make a staging system simple. We want to be accurate, but simple, and it wasn't easy to reconcile both forces.

Now, for patients with primary melanoma, level of invasion was relevant only for thin lesions, less than a millimeter.

For everything above one millimeter, ulceration was the second most powerful factor. So, Clark was not included except for thin lesions.

Slide 17:

In patients in whom we had the histologic status of the regional node, either by regional dissection or sentinel node biopsies -- about 5,000 patients -- knowing the histologic status of the node was the most powerful prognostic factor.

Thickness and ulceration were also important, but obviously the histologic status was the most powerful one.

Slide 18:

In fact, when you look at various categories, patients with obviously node positive by sentinel node biopsy was important for every category.

In no category the survival curve will go down.

Slide 19:

The only exception -- it is interesting -- is in patients with more than four millimeters and ulcerated, T4b, at the very bottom.

It was so bad, the primary, that knowing the histologic status of the node did not make a lot of difference.

Slide 20:

Once you have a patient with nodal metastases established with palpable nodes, then the most important factor is the number of positive nodes.

The second-most important factor was whether it was microscopically affected or macroscopically affected.

An illustration of interest, in the primary is to affect the prognosis of patients with nodal metastases.

Slide 21:

This is just to illustrate what everybody knows. Patients with one positive node, up to greater than four, the survival curves will go down. This has been shown for years in every data base.

Slide 22:

Of interest, however, is this stratification. If you get a patient with a single positive node, and the patient has microscopic nodal involvement, or macroscopic nodal involvement, 61 percent are alive at five years if they have only microscopic nodal involvement, compared to 46 for macroscopic. This goes down for every single category.

Slide 23:

Of interest, this is also true for ulceration, indicating that ulceration indicates a more aggressive biologic behavior.

Now, as an example here, one positive node, if the primary is ulcerated, there is a 45 percent five-year survival. If the primary is not ulcerated, a 65 percent.

Slide 24:

Now, patients with regional nodal metastases like satellites or intransit lesions behave in a similar manner to patients with nodal metastases.

Slide 25:

The distant disease category, I don't think we did any service to this group, in my opinion, in the new staging system, based on this data, which is relatively small.

Patients with lung metastases fell a little bit in the middle. It starts together with skin and finished with visceral, and I don't think we did a really good job here.

Slide 26:

LDH was already mentioned, I think initially described by Don Morton many years ago. I was born at that time, but it was a while back.

The patients with high LDH obviously do more poorly than patients with low LDH. That is why it was included.

Now, I am not going to go over all the groupings because it is laborious and not worth it, but just to emphasize that now, for the T category, the TNM staging system,

Slide 27:

the cuts offs are better, they are more powerful and they are simpler to remember, one, one to two, two to four, and greater than four.

Ulceration is the primary determinant for A and B category, and Clark is relevant only for thin lesions. Patients with Clark 4 are considered as if they had ulcerated thin lesions.

Slide 28:

In the nodal category, important now is the number of positive nodes. The second determinant is now whether it is microscopically involved, usually detected by sentinel node technique, or macroscopically involved.

Patients with poor features, such as more than four positive nodes, or matted nodes, or those with intransit and nodal involvement, fall into the N3 category.

Slide 29:

Finally, the stage four, the M category, basically the only thing of relevance here is the fact that if you have a high LDH, you have a worse prognosis and you fall into an M1C category.

Slide 30:

The curves separate well when you plot by stage, and as we would expect, in fact, it was designed that way, obviously.

Slide 31:

Where I would like to take you, I used to see only melanoma, almost, the last 15 years of my life. I now see other cancers, a lot easier to handle, in my opinion.

I see breast cancer and I see testes cancer. I am even embarrassed I have to charge the patient. The point is, there are certain lessons from other cancers that we should bring to the melanoma field. There are two important papers in breast cancer that we should definitely review critically.

Slide 32:

One is using microarray technology. Plotting on the left is node negative patients and on the right is node positive. Based on the gene profile, they called it good signature or bad signature.
Now, look on the right side. This is the node positive patients.

There is not a single series that have node positive patients with greater than 90 percent overall survival.

So, they are really able to pick an incredible group of patients. This was relevant by micro array technology. It is from Holland.

Slide 33:

The second paper is from M.D. Anderson. So, this is very impressive. If you look at this curve, I divide the curves into provocative, interesting and almost too good to be true. This fell in the third one.

This is cyclin expression by western blot in breast cancer, stage one through three. If I were to cheat something and tell you, pretend you just found the most incredible prognostic factor, I don't think I would have put it that wide.

Now, the separation is very impressive and I wonder, as cancers tend to spread by similar means, I wonder if this couldn't be true for melanoma, for instance. It is remarkable that this could be only for breast cancer.
Now, it has to be validated, obviously. If this is validated, in my opinion, breast cancer research would be incredibly modified in terms of stratifying patients in adjuvant trials.

Slide 34:

Well, how about the future? That goes to the eight-page memo from Vern.

Slide 35:

[No text is associated with this slide].

Slide 36:

I don't think it is going to be that easy to say how the future of the staging system will be.

Let's look back at Bodenham's 1968 paper that divides the melanoma in the good and bad, a very simple staging system.
What happened, since 1968 we became very complex. The new staging system has a tremendous number of subsets.

Even I have to look there sometimes to verify if everything is correct.

Now, how will be the future, however? The future, in my opinion, will be very simple. It will be called the A and D staging system, A for alive and D for dead.
I don't know at which ASCO meeting will be presented, but it will be exactly that way. In fact, a report I dare to show you--

Slide 37:

--the report will be like this.
Based on a gene expression profile by microarray technology, this patient has stage A or, based on the gene expression profile this patient has, without any therapeutic intervention, stage D.

However, the gene expression profile indicates that with the use of the following drugs, this patient has a cure rate in the range of 90 percent. That is the way I foresee the future.

Slide 38:

Now, for people that asked, what happened to me? I did lose some weight while at Yale, but I did gain a lot of muscle at Anderson, and I am not doing that badly.

I am back in Brazil. Those are assistants who are happily married, and I will finish up there. Thank you very much.