[No text is associated with this slide]

TOP

I have the daunting charge of comparing and contrasting familial and sporadic melanoma, from Dr. Sondak. I will try to do what I can in a short period of time.

Contrary to rumors that may be floating around, melanoma is not decreasing in this country. It is still rapidly increasing.

The most rapid increase -- these are the most recent data from SEER. If you look at the most recent published data down here, it looks as though it might be decreasing.

What has happened in SEER is that there is a lag between final reporting and initial reporting, because more melanomas are being taken off not in hospitals.
If you look at what the final reporting is, melanoma is still rapidly increasing in this country, both in men and women. Obviously, the rates are higher in men.

TOP

As Scott told you, there are expected to be 54,000 new cases of melanoma this year. It is the fifth most common cancer in men, accounting for four percent of cancers in men, and the seventh most common in women, accounting for three percent.

The lifetime risk has increased to 1.75 for men and 1.23 for women. There are certain subgroups of the population in which melanoma is increasing more rapidly. That is in men over age 50 and women in their 20s and 30s.

TOP

As I said, this is total melanoma, and some of you have often wondered why we don't have more tissue to play with for melanoma. This is the reason.

These also are based on SEER data. If you look at the number of melanomas diagnosed in all of the United States, with regional disease at the time of diagnosis, it is close to 5,000, and metastatic is about 2,000.

If you are talking about tumors that are large enough that the pathologist doesn't need all of it, clearly that is true for four or more millimeter thick melanomas, but that is only 2,300 in the entire United States.

If you talk about the vast majority of these, they are actually one to two millimeters, and it is difficult for a pathologist to give that up. So, you are really talking about fairly limited biologic specimens to be doing work with directly.

TOP

Moving to families, these are essentially in order of risk. Xeroderma pigmentosum (XP) obviously has about a thousand-fold increased risk, but these families are vanishingly rare.

Familial melanoma is what I am going to be spending most of the time talking about. You will hear more about that.

Retinoblastoma confers roughly 50 to 100-fold increased risk of melanoma. Retinoblastoma, heritable retinoblastoma, individuals have about a 50-fold increase of melanoma.

Werner syndrome and possibly Li-Fraumeni are at increased risk of developing melanoma. These two are vanishingly rare, as are retinoblastoma and XP.

So, what we are really talking about is familial melanoma. Even within this relatively restricted group of familial melanoma, the disease is complex and heterogeneous.

TOP

There are essentially two definitions of familial melanoma. One is two or more first degree relatives.

If you ask newly diagnosed melanoma cases, eight to 10 percent of those individuals will say they have a first degree relative with melanoma.

This level of familial risk confers about a two-fold increased risk of melanoma overall, in virtually every epidemiologic study that has evaluated it.

With three or more living relatives, that restricts the definition significantly, and that accounts for about one percent of melanomas.

These are the type of families that are used for linkage analyses, and these families have a much, much higher risk.

TOP

If you look at familial melanoma or sporadic melanoma under the microscope, they look very similar.

These are other characteristics. Family history, obviously, in sporadic, they have none. Familial, we just went over.

The age at diagnosis, the average age of diagnosis in sporadic in SEER is 53 years old, in this country. Familial is 35 years old.

Number of primaries in SEER, 98 percent of individuals with melanoma only develop one primary melanoma, whereas 40 percent of individuals in the high risk linkage type families develop multiple melanomas.

In many epidemiological studies, about half of individuals who develop melanoma have dysplastic nevi in this country, and most of the familial ones do.
The lesions tend to be somewhat thicker in sporadic and familial are thinner because they frequently are enrolled in research protocols and are under surveillance.

TOP

The major susceptibility genes for melanoma are well known to you all. CDKN2A is a tumor suppressor gene and, really, is the major susceptibility gene for melanoma.

There are two transcripts, p16 and p14ARF. Most of the mutations that affect CDKN2A affect p16, and the majority of those also affect the function of p14ARF.

There is a small handful of families that have been reported with mutations that affect p14RAF only. Many of the mutations that are recurrent in CDKN2A are founder mutations, and those mutations date back 30 to 100 generations, depending on the specific mutation.

Relatively restrictive populations that have looked for founder mutations have found that the founder mutations that initiated in Europe have spread to Canada, the United States and Australia.

CDK4 is an oncogene and there have been three families reported in the literature so far. There are probably a couple of others who have been found that have not yet appeared in the literature.

This is extremely rare. There are others actively being sought. The international melanoma genetics consortium is working with Jeff Trent to look for new high risk susceptibility genes, but it is likely that p16 is still going to be the major player.

TOP

The probability of finding a CDKN2A mutation depends on the family history. In families with only two affected first degree relatives, the chance of finding a mutation is less than five percent.

If you talk about three or more affected first degree relatives, if you average all of them together, it will turn out to be about 20 percent, because many of those will be at three relatives.

If you talk about very, very loaded families, you get probably up over 50 percent, somewhere around 60 percent, will have a p16 mutation.

If you look at individuals who present with multiple primary melanomas, without a family history, about 10 percent will have a p16 mutation.
If you look at all comers, it is about 10 to 15 percent of individuals with multiple primaries will have a CDKN2A mutation.

TOP

The prospective risk of melanoma, these are data from our families. The prospective risk of melanoma in all of our families is a 35 to 70-fold increase, independent of type of mutation that they have.

The major risk factors that have been identified in our families are germ line mutations in CDKN2A or CDK4, dysplastic nevi and sun exposure. Both dysplastic nevi and sun exposure alter the risk of melanoma in mutation carriers in our families.

TOP

Dysplastic nevi occur equally in our families with different types of mutations and they clinically are indistinguishable among members of families with different types of mutations.

This is a dysplastic nevus on someone with a CDKN2A mutation, one on someone with a CDK4 mutation, and this is a dysplastic nevus on an individual with heritable retinoblastoma who also had melanoma.

TOP

We have been following families for up to 25 years. In those 25 years, a number of individuals have developed new melanomas.

These are data from a recent article of ours. Essentially, looking at prospective melanomas, seven percent of them are in this group that are over one millimeter thick.

In fact, the risk of melanoma, the number of multiple primaries, does not differ between families with CDK4, CDKN2A and as yet unidentified mutations.

TOP

In CDKN2A mutation carriers, dysplastic nevi confer an eight-fold increased risk. Total nevi, which is indistinguishable from dysplastic nevi in this data set, also confer risk and solar injury about the level in population.

TOP

To estimate the penetrance or risk of developing melanoma in CDKN2A carrier, the International Melanoma Genetics Consortium collaborated to evaluate melanoma status in 80 families from the United States, Australia and Europe.

The method of analysis was a logistic regression incorporating survival analyses, and specific covariates were examined to see if they altered the risk of melanoma in CDKN2A mutation carriers.

This included gender mutations affecting p14ARF and population rates of melanoma.
In the multivariate analyses, the only thing that altered risk significantly were the population rates of melanoma.

TOP

Here are the data. This appeared in JNCI in I think July last year. These are the overall rates going to a high of 67 percent overall by age 80.

This is Europe, Australia and the United States. You can see the shapes of the curve is different and the risk at age 50 is quite different, being only 13 percent in Europe and actually 58 percent in the United States. The cumulative rates to age 80 are also quite different.

TOP

So, based on these penetrance findings, we reassessed the 1999 recommendations for the utility of genetic testing in these high risk families.

The general consensus was that it was premature to offer CDKN2A testing in most circumstances at this time because, as I reviewed with you, the likelihood of finding a mutation is low in most families. The penetrance, as I just showed you, is highly variable.

Most important, knowing mutation status does not change recommended clinical care for family members already under surveillance and that are following guidelines that have been developed over the years.

TOP

So, are there substantial differences between CDKN2A and CDK4? Alicia Goldstein looked at this in our families and, essentially, the age of onset is virtually identical.

The number of new primary melanomas is virtually identical, and the number of nevi is virtually identical. These are based on small numbers and need to be replicated with much larger numbers, which we hope to do within the genetics consortium.

TOP

Obviously, there are a lot of other genes that are of intense interest and are being evaluated all over the world.

These include pigmentation genes, MC1R, agouti pathway, DNA repair genes, especially those involved in UV-induced DNA repair, retinoblastoma in a limited manner, immunomodulatory genes, because melanoma is increased in those folks who are immunosuoppressed, and that is clearly one of the ways that we can treat melanoma, because of Werner's helicase, and people are actively looking for potential nevus genes as modifier genes.

TOP

Going back, to remind you again, this is total United States.

TOP

If we are talking about families, these are the numbers for the United States.

Every year, there are potentially -- if you assume that it is one percent, that is 500 new melanomas in the United States.

About 100 will have CDKN2A mutations and only 40 of the 540 will be over a millimeter thick, if the same thickness rates pertain as we have.

TOP

Basically, these are data from our case control study, which are similar to most epidemiological studies, identifying risk factors. Even if we don't use mutation status, we can identify high risk individuals.

TOP

If you look at the mutation carriers, that is about 100 people per year. They have, again, about a 50-fold increased risk of melanoma.

If you take people with less than one millimeter melanomas, invasive melanomas, that is going to be about 40,000 people.

They are already enrolled in the melanoma surveillance system. They are identified by the folks who are interested in melanoma.

Within the first year, they have a 15-fold increased risk of melanoma.
These are SEER rates and, within the first five years, a 10-fold increased risk of melanoma, clearly a good population to look at.

Multiple dysplastic nevi are about five percent of the U.S. population, and they have about a five to 10-fold increased risk of melanoma.

TOP

So, the melanoma susceptibility genes are extremely rare. It is clear that retinoblastoma is a critical pathway in melanoma development.

Identifying these genes has been essential in understanding the biology of the development of melanoma and potentially therapy but, at this point, it is not a basis for clinical decision making.

There are other high risk individuals, however, that can be clearly identified, and there are well characterized risks within those populations.

TOP