- Top of page
- MATERIAL AND METHODS
Familial atypical multiple mole melanoma (FAMMM) is an autosomal dominant disease characterized by the familial occurrence of malignant melanoma of the skin and multiple atypical precursor lesions. Germline mutations in the p16 (CDKN2A) gene have been reported in at least a quarter of such families. An association has been reported between p16 mutations and pancreatic cancer. The aim of this study was to assess the risk of developing pancreatic and other cancers in Dutch FAMMM families with a 19 bp deletion in exon 2 of the p16 gene (p16-Leiden). Mutation analysis was performed in 27 families suspected of FAMMM. Clinical and pathological data were collected from all relatives affected with cancer. A p16-Leiden mutation was identified in 19 families. These families included 86 patients with melanoma. The second most frequent cancer was pancreatic cancer, which was observed in 15 patients from 7 families. The mean age at diagnosis of pancreatic cancer was 58 years (range 38–77 years). The estimated cumulative risk of developing pancreatic cancer in putative mutation carriers by age 75 years was 17%. In 8 p16-Leiden-negative families, no cases of pancreatic cancer occurred. p16 mutation carriers have a considerable risk of developing pancreatic cancer. Further studies should evaluate the value of surveillance of the pancreas in these high-risk families. Int. J. Cancer 87:809–811, 2000. © 2000 Wiley-Liss, Inc.
Familial atypical multiple mole melanoma (FAMMM or familial dysplastic nevus syndrome) is characterized by the familial occurrence of malignant melanoma of the skin in combination with multiple atypical precursor nevi (Greene et al.,1987). The disease is inherited as an autosomal dominant trait. In 1989, linkage was reported with a gene locus on chromosome 1p36 (Bale et al.,1989); however, this finding was not confirmed in other families. Three years later, definitive linkage was obtained with a locus on chromosome 9p21 (Cannon-Albright et al.,1992). This locus encompasses the p16 (CDKN2A) gene. This cyclin-dependent kinase inhibitor is probably responsible for melanoma susceptibility in at least a quarter of all FAMMM families. Another gene reported to be involved in some rare FAMMM families is CDK4. The penetrance of the disease gene reported in the literature (the likelihood that an individual develops melanoma by age 80) varies from 53% to 100% (Cannon-Albright et al.,1994; Walker et al.,1995). Surveillance programs including regular examination of the skin from the age of about 12 years have been recommended in these families. At least 2 studies have reported that such programs led to the detection of melanoma at an earlier stage (Vasen et al.,1989; Masri et al.,1990). Studies from the early 1990s suggested that patients are at risk not only for melanoma but also for developing other cancers, particularly pancreatic cancer (Bergman et al.,1990). The increased risk of pancreatic cancer in these families has so far been observed only in families associated with germline mutations in the p16 gene (Goldstein et al.,1995). A retrospective follow-up study going back to 1830 on 6 FAMMM families associated with p16-Leiden mutations revealed that in some families the mortality due to pancreatic cancer exceeded the mortality associated with melanoma (Hille et al.,1998). The aims of the present study were to evaluate the risk of developing pancreatic and other cancers in FAMMM families registered at the Netherlands Foundation for the Detection of Hereditary Tumours and to discuss possible secondary preventive measures.
- Top of page
- MATERIAL AND METHODS
The 19 families with a p16-Leiden mutation included a total of 656 relatives. Table I shows the number of patients with melanoma or other tumors in the families with a p16-Leiden mutation and those without a p16-Leiden mutation. Mean ages at diagnosis of the first melanoma identified in families with and those without p16-Leiden mutation were 39 years (range 15–72) and 44 years (12–71), respectively. Families with a p16-Leiden mutation exhibited, on average, 4.5 melanomas per family, while those without a p16-Leiden mutation exhibited 3 melanomas per family.
Table I. PREVALENCE OF CANCER IN 27 FAMMM FAMILIES TESTED FOR p16 MUTATIONS
|Cancer site||Families with p16 mutation (n = 19)||Families without p16 mutation (n = 8)|
|Melanoma||86||24 (including 4 eye melanomas)|
|Other||9||Morbus Kahler (n = 1)|
| Prostate cancer (n = 1)|
p16-Leiden mutations were identified in 4 (50%) of the 8 families with 1 or 2 melanomas per family, in 6 (75%) of 8 families with 3 or 4 melanomas, in 4 (66%) of 6 families with 5 or 6 melanomas, and in all 5 (100%) families with 7 or more melanomas.
Pancreatic cancer was the most frequent type of cancer in the p16-Leiden families after melanoma, being observed in 7 of 19 families. Two families included 4 cases of pancreatic cancer, 2 families included 2 cases, and the remaining 3 families included 1 case each.
The mean age at diagnosis was 58 years (range 38–77). Detailed information on these cancers identified in families with a mutation is shown in Table II. Table II also includes information on 12 cases of pancreatic cancer in other FAMMM families known at the FAMMM Registry which have not been tested for mutations as they did not fulfill the inclusion criteria. The mean age at diagnosis of pancreatic cancer in the families with a mutation (59 years) was not different from the mean age at diagnosis of cancers (58 years) identified in families that were not tested. Table III shows the cumulative risk of developing pancreatic cancer by age for the (putative) mutation carriers (i.e., proven mutation carriers and patients with a melanoma) and their first-degree relatives.
Table II. CHARACTERISTICS OF PANCREATIC CANCER IN p16-POSITIVE AND OTHER FAMMM FAMILIES
|Number||Family ID||Sex||Age (years)||Basis for diagnosis||p16 mutation|
Table III. AGE-SPECIFIC RISK OF DEVELOPING PANCREATIC CANCER (%) IN p16-LEIDEN-MUTATED FAMMM FAMILIES
|Age (years)||Carriers (putative)1||First-degree relatives2|
|753||17 (95% CI 3–30)||13 (95% CI 2–24)|
In addition, a relatively high number of cases of lung cancer was observed (11 cases in the mutation carriers). Lung cancer was diagnosed in 8 families. Three families comprised 2 cases and 5 families included 1 case each. The cumulative risk of lung cancer by age 75 years in male (putative) carriers of a p16-Leiden mutation was 14.3% [95% confidence interval (CI) 1.4–27.2] and in female carriers it was 1.9% (95% CI 0–5.7). The expected cumulative risk of lung cancer by age 75 years in the Netherlands is 8.9% for men and 1.65% for women (Visser et al.,1993). Both pancreatic cancer and lung cancer occurred in 4 of the 19 families.
- Top of page
- MATERIAL AND METHODS
A retrospective follow-up study was conducted in families known at the Dutch FAMMM registry. Mutation analysis was performed in families with at least 2 first-degree relatives with a melanoma. Twenty-seven families fulfilled these criteria, and in 19 families a specific 19 bp p16 mutation, referred to as p16-Leiden, was identified. Pancreatic cancer was found only in p16-Leiden-positive families. In an earlier report (Bergman et al.,1990) from our group, pancreatic cancers were observed in 3 of 9 thoroughly investigated FAMMM families.
On the basis of our present results, we estimate that 1 of 6 (17%) p16-Leiden mutation carriers will develop pancreatic cancer by age 75 and ultimately die from this disease. Interestingly, pancreatic cancer was observed in only some (7 of 19 families) of the families with this specific mutation, which suggests that members of families that include cases of pancreatic cancer have an even higher risk of developing this cancer than those in families without pancreatic cancer. Possible explanations for this variation in expression are the presence of modifying genes or the influence of environmental factors (e.g., smoking or nutrition).
As the present study is retrospective, the results might be biased. For example, families with cases of pancreatic cancer might have been selected for DNA analysis just because of the reported association between this cancer and the p16 gene. However, molecular genetic analysis of the families was started before this association was known. In the first 6 families, extensive genetic testing of relatives was performed, while in the remaining families, only 1 or 2 melanoma patients were investigated. All patients with melanoma observed in the latter families were considered to be (putative) mutation carriers. Interestingly, in the first 6 families, 4 patients with melanoma were found not to carry a mutation. This means that melanoma patients in the other 13 families might be non-carriers. As these patients were considered in the calculations to be mutation carriers, the risk of developing pancreatic cancer might be under-estimated.
Pancreatic cancer has also been reported in other inherited tumor syndromes, including hereditary non-polyposis colorectal cancer (Lynch et al.,1985), Peutz-Jeghers syndrome (Giardiello et al.,1987), and familial breast cancer associated with BRCA2 mutations. Also, families with exclusively pancreatic cancer have been reported. It is questionable whether families with site-specific pancreatic cancer have an underlying p16 mutation.
Knowledge of the specific features of pancreatic cancer associated with FAMMM is important for the management of this cancer. Hereditary cancer is generally associated with an unusually young age at diagnosis and the occurrence of multiple tumors. However, the mean age at diagnosis of pancreatic cancer in site-specific pancreatic cancer families known at the National Familial Pancreas Tumor Registry (Hruban et al.,1999) was not different from the age at diagnosis of unselected cases of pancreatic cancer. The average age at diagnosis in both groups was 65 years. In our FAMMM families, the age at diagnosis of pancreatic cancer was slightly lower (59 years).
There is evidence that surveillance of FAMMM families leads to early detection of melanoma. In contrast, it is generally recognized that screening for pancreatic cancer is extremely difficult. Only reports from Japan suggest that screening by abdominal ultrasound of average-risk individuals led to early detection of pancreatic cancer (Ariyama et al.,1998). The effectiveness of screening high-risk groups such as FAMMM families is unknown. The screening protocol suggested in the literature (Lynch et al.,1996; Brentnall et al.,1999) for high-risk families includes annual estimation of alkaline phosphatase, amylase, lipase, carcinoembryonic antigen, and CA19.9. Proposed imaging techniques include abdominal ultrasound, computed tomography, endoscopic retrograde cholangiopancreaticography, endoscopic sonography, and positron emission tomography. In our opinion, surveillance of FAMMM families for pancreatic cancer should be performed only in the setting of a prospective study. In such studies also, the value of new screening methods, such as estimation of K-ras oncogene in the feces or bile, might be assessed. On the basis of the distribution of age at diagnosis of the pancreatic cancers in the present series, we recommend starting such a screening program at an age between 45 and 50 years.