Introduction

The major categories of skin cancer are malignant melanoma and non-melanoma skin cancers (keratinocyte carcinoma) which include basal cell carcinomas and squamous cell carcinomas [1]. The World Health Organization estimates that 1,200,000 cases of non-melanoma skin cancer and 325,000 cases of melanoma were diagnosed in 2020 (with 64,000 deaths and 57,000 deaths respectively) [2].

Risk factors for skin cancer include sun exposure (ultraviolet light), fair skin and eye colour, and a family history of skin cancer [3]. Germline mutations in several genes have also been associated with skin cancer predisposition. Variants in SUFU and PTCH are responsible for the nevoid basal cell carcinoma syndrome (NBCCS) also known as Gorlin syndrome [4, 5]. Variants in CDKN2A are responsible for Familial Atypical Multiple Mole Melanoma syndrome (FAMMM), which accounts for almost 40% of families with multiple cases of cutaneous melanoma [6, 7]. Other genes associated with the development of melanoma include PTEN, CDK4, ACD, TERF2IP BAP1, POT1, MITF, MC1R, TP53, RB1 and TERT [8,9,10,11,12,13,14,15]. Together, the known predisposition genes account for approximately 50% of familial melanoma cases [9] but fewer than 10% of all melanoma cases.

Women who carry a BRCA1 or BRCA2 mutation have elevated risks for many types of cancer, including breast, ovarian, fallopian tube and peritoneal cancers [16]. Men are at increased risk for prostate cancer and both sexes are at risk for pancreatic cancer. The evidence for an increased risk of skin cancer is inconsistent. Early studies of women who carry BRCA mutations found an increased risk of melanoma in family members of BRCA2 carriers [17, 18] However, this was not seen in a recent study of BRCA1 and BRCA2 families described by the CIMBA consortium [19]. Family based studies may be subject to bias and all relatives are not known to be carriers. A better approach is to follow a cohort of BRCA carriers over time and document new cases of skin cancer in established carriers. We sought to estimate the annual and lifetime risks for skin cancer associated with BRCA mutations in order to determine if screening measures should be adopted.

Methods

Study participants were identified from a multi-centre longitudinal study of women who carry BRCA1 or BRCA2 mutations (pathogenic variants). Women were drawn from participants in ongoing studies at 52 participating centres throughout Canada and the United States. All centres received approval by their Research Ethics Board or Institutional Review Board. Each participant provided informed consent and filled out a baseline questionnaire at study entry and a follow-up questionnaire every two years thereafter. These questionnaires included information about personal history of cancer including skin cancer, prior cancer diagnoses, family history of cancer, and ethnicity Women were included if they had completed at least one follow-up questionnaire and were excluded if they were lost to follow-up or were missing BRCA mutation status. Women with a previous diagnosis of cancer, including breast, ovarian, skin or other cancer at study entry were included. Family history of skin cancer in a first-degree relative was recorded when available. The majority of the participants were white of European ancestry.

Participants were asked in each follow-up questionnaire if they had been diagnosed with any cancer using an open-ended question. Skin cancer diagnoses were based on patient report and fell into three categories: basal cell carcinoma, melanoma or ‘skin cancer’. Women were followed from the date of their baseline questionnaire until diagnosis of skin cancer, date of last follow-up questionnaire or death.

Statistical analysis

The annual risks for skin cancer (all types) and for melanoma were estimated by age group (20–29, 30–39, 40–49, 50–59, 60–69, and 70–79 years) for BRCA1 and BRCA2 carriers by calculating the ratio of number of events (cancer diagnoses) and the number of person years in each group. Person years were accumulated in the cohort for each age group (e.g. age 20–29 years) for those women who were alive and cancer-free at the beginning of each age group (e.g. age 20 years). The number of skin cancers that developed in women in that age group were then counted.

The cumulative incidence of skin cancer (any) and of melanoma was calculated using the Kaplan-Meier method. Patients were followed from age 20 or baseline (whichever came last) until the diagnosis of skin cancer, death from any cause, age 80 or the date of the last follow up.

The univariate and multivariate hazard ratios (HR) and 95% confidence intervals (CIs) associated with various risk factors for skin cancer were estimated using the Cox Proportional Hazards model. Exposures of interest included a family history of skin cancer, a personal history of skin cancer, past history of breast cancer and hormone replacement therapy. Subgroup analysis was performed by mutation status (BRCA1, BRCA2) and by type of skin cancer (all skin cancer, melanoma). All analyses were conducted using the SAS statistical package, version 9.4 software (SAS Institute, Cary, NC, USA). All P values were two-sided and were considered statistically significant if p < 0.05.

Results

A total of 6,207 women were included in the analysis: 3,623 women with BRCA1 mutations and 2,594 women with BRCA2 mutations (see Table 1). The average duration of follow-up from baseline was eight years. A previous diagnosis of breast cancer was reported by 45.5% of participants and a previous diagnosis of ovarian cancer was reported by 10.5%. A previous diagnosis of skin cancer was reported by 3.4% of BRCA1 mutation carriers and by 4.1% of BRCA2 mutation carriers. 173 women reported a family history of skin cancer in a first-degree relative (6.7%).

Table 1 Characteristics of the 6207 subjects
Full size table

During the follow-up period, 232 women reported a diagnosis of skin cancer (3.7%); 88 basal cell carcinomas, 44 melanomas and 100 ‘skin cancers’. The age-specific annual risks for skin cancer (all types combined) and for melanoma are presented in Tables 2a and 2b. Overall, between age 40 and 79 the annual risk for any skin cancer was 0.5% for BRCA1 carriers and 0.6% for BRCA2 carriers. Between age 40 and 79 the annual risk for melanoma was 0.09% for BRCA1 carriers (90/100,000 per year) and 0.12% for BRCA2 carriers (120 per 100,000 per year). We estimated the cumulative risk of any skin cancer from age 20 to 80 to be 14.1% for BRCA1 mutation carriers and 10.7% for BRCA2 mutation carriers. We estimated the cumulative risk of melanoma between age 20 and 80 years to be 2.5% for women who carry BRCA1 mutations and 2.3% for women who carry BRCA2 mutations. Women from the United States had a higher risk of skin cancer than those from Canada (HR = 1.98, 95% CI: 1.39, 2.83, p < 0.0001) for BRCA1 carriers and HR = 1.55 (95%CI 1.04–2.30; p = 0.03) for BRCA2 carriers (Tables3a and 3b).

Table 2a Annual risks for skin cancer (all types) by age and gene
Full size table
Table 2b Annual risks for melanoma by age and gene
Full size table
Table 3a Hazard ratios for skin cancer for some variables (BRCA1 subjects)
Full size table
Table 3b Hazard ratios for skin cancer for some variables (BRCA2 subjects)
Full size table

A diagnosis of skin cancer prior to study entry was the strongest risk factor for development of skin cancer (all types) (Tables 3a and 3b). The univariate hazard ratio was similar for BRCA1 carriers (HR = 3.84, 95% CI: 2.24–6.58, p < 0.001) and for BRCA2 carriers (HR = 2.80, 95% CI: 1.50–5.24, p = 0.001).

Overall, four of the 44 women with a new diagnosis of melanoma had a prior diagnosis of melanoma. Women who reported a diagnosis of melanoma prior to study entry had a significantly elevated risk for a new diagnosis of melanoma (Tables 4a and 4b). The hazard ratio was 11.9 (95% CI: 2.78–50.9; p = 0.0008) for BRCA1 carriers and was 9.90 (95% CI 2.28–4.27; p = 0.002) for BRCA2 carriers (see Table 4a and 4b). None of the women with incident melanoma had a family history of melanoma.

Table 4a Hazard ratios for melanoma for some variables (BRCA1 subjects)
Full size table
Table 4b Hazard ratios for melanoma for some variables (BRCA2 subjects)
Full size table

Discussion

We estimated the annual and cumulative risk of skin cancer in women who carry BRCA1 or BRCA2 mutations. We found the risk for melanoma in BRCA1 and BRCA2 mutation carriers to be only slightly higher than expected, based on US population rates. The lifetime risk of melanoma was approximately 2.5% for women who carried BRCA1 mutations and 2.1% for women who carried BRCA2 mutations, compared to 1.5% for white women in the general population in the United States [20]. Similar data are not available for non-melanoma skin cancer. Moran et al. reported two cases of uveal melanoma in 222 families with BRCA2 mutations (HR = 99.4; 95%:CI 11.1–359) [18]. In that study the risk was borderline significant for skin melanomas among BRCA2 carriers (6 observed, 2.3 expected); p = 0.05). There was incident case of uveal melanoma in a 49 year old women with a BRCA1 mutation in our study who was not included in the analysis.

A prior history of skin cancer was the strongest risk factor for a new skin cancer. We estimated a hazard ratio of 9.90 for a second diagnosis of melanoma after initial diagnosis in BRCA2 mutation carriers and 11.4 for those with a BRCA1 mutation. This is consistent with data from the general population: up to 10% of melanoma patients will go on to develop a second melanoma [21, 22]. NCCN provides recommendations for increased follow-up and screening after initial diagnosis of melanoma followed by annual skin exams with education on self-examination [23]. NCCN states that regular clinical examination has the highest diagnostic benefit and is the most cost effective method for early detection of treatable disease [23].

The small increase in melanoma risk for BRCA1 and BRCA2 mutation carriers is consistent with the findings of an early study of cancer risks among a cohort of women who carry BRCA2 mutations and their first-degree relatives compared with population-specific incidence rates which found a relative risk of melanoma of 2.6; (95% CI = 1.3–5.2)24. Mersch et al. found an excess of a history of melanomas in a sample of mutation carriers at MD Anderson hospital [24]. Tuominen et al. also found an association with a founder mutation in a case control study of BRCA2 in Swedish melanoma patients with an odds ratio of 2.80 (95% CI: 1.04–7.58, p = 0.035).[25] Debniak et al. did not find and excess of BRCA1 or BRCA2 mutations in a sample of melanoma patients from Poland [26].

In an earlier study of this cohort, we reported an excess risk of basal cell carcinoma in BRCA1 and BRCA2 carriers [27]. The current study expands the cohort size from 2797 women to 6207 women and extended the mean follow-up from 5.0 years to 8.1 years. To our knowledge, this is the largest prospective study addressing the risk of skin cancer in BRCA mutation carriers. We found the cumulative risk of all types of skin cancer reported in this cohort from age 20 to 80 years to be 14.1% for BRCA1 mutation carriers and 10.7% for BRCA2 mutation carriers. In comparison, the lifetime risk for basal cell carcinoma and squamous cell carcinoma in women in the general United States population has been estimated to be approximately 28–33% and 7–11% respectively for a woman born in 1994[28]. While these data suggest that the risk for skin cancer in BRCA mutation carriers is not increased, data on the incidence of non-melanoma skin cancer is limited as it is not included in most cancer registries [29, 30]. Our cohort has a mean birth year of 1957 which also bears on the comparison of these risks. Studies have indicated that the incidence of keratinocyte carcinoma has been steadily increasing [30].

We found that the risk for skin cancer was increased after a diagnosis of breast cancer for BRCA1 mutation carriers (HR = 1.77, 95% CI: 1.26–2.50, p = 0.001) but not for BRCA2 mutation carriers (HR = 1.13, 95% CI: 0.76–1.67, p = 0.56). A relationship between breast cancer and skin cancer has been observed in cancer registry-based studies [31,32,33]. A large study looking at second primary cancers after breast cancer in the general population found a similar association with SIRs for non-melanoma skin cancer and melanoma of 1.58 and 1.29 respectively [32].

There are a number of limitations in this study. This study of skin cancer risk was based on patient report of new cancer diagnoses in biannual questionnaires and diagnoses were not confirmed with pathology reports. Women with incident skin cancers reported either melanoma, basal cell carcinoma, or ‘skin cancer’ and it is possible that the sub-types of cancers were misclassified. It is possible that skin cancer is under-reported by study participants as well as in their first-degree relatives. We cannot estimate the risks for male BRCA mutation carriers. These data are derived from predominantly white women in Canada and the United States. It is possible that the risks are higher in white women in countries with greater sun exposure, such as Australia and New Zealand. We have information on current residence, but not prior residence and it is possible that early childhood exposures are relevant for skin cancer risks in adults. We do not have information on other risk factors for skin cancer such as hair color and fair skin and extent or recreational sun exposure. Our cohort was followed until a mean age of 55 years and skin cancer rates were highest above age 70. Future studies will be conducted to better estimate the lifetime cumulative rate of skin cancer among BRCA1 and BRCA2 carriers. Further, the results of this study should not be extrapolated to non-white women and there were too few non-white women in the study to conduct a robust analysis.

Prevention of skin cancer through limiting ultraviolet light (UV) exposure, use of sunscreen and early detection is important in the general population. The NCCN states that no specific screening guidelines exist for melanoma for BRCA mutation carriers, but that general melanoma risk management such as annual full-body skin examination and minimization of UV exposure is appropriate [34]. The results of our study are consistent with conclusions of the review article by Gumaste et al. [35] who stated there was insufficient evidence to support the claim that skin cancer is a manifestation of the BRCA1 and BRCA2 phenotype but recommend annual skin examinations. We suggest a referral for skin exam to a primary care provider for BRCA mutation carriers with a prior history of skin cancer for annual skin examination and counselling around protection from UV exposure, use of sunscreen, and self-screening to recognize the early signs of melanoma.