|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 24, No 19 (July 1), 2006: pp. 3172-3177 © 2006 American Society of Clinical Oncology. DOI: 10.1200/JCO.2006.06.1325 Anatomic Site, Sun Exposure, and Risk of Cutaneous Melanoma
From the Queensland Institute of Medical Research, Brisbane, Australia Address reprint requests to David C. Whiteman, MBBS, PhD, Division of Population Studies and Human Genetics, Queensland Institute of Medical Research, PO Royal Brisbane Hospital, Queensland 4029 Australia; e-mail: david.whiteman{at}qimr.edu.au
PURPOSE: Sunlight is the principal environmental risk factor for cutaneous melanoma. A current hypothesis postulates that the role of sunlight in causing melanoma differs according to anatomic site. We tested this hypothesis in a population-based case-case comparative study of melanoma patients. METHODS: Patients were sampled from the Queensland cancer registry in three groups: superficial spreading or nodular melanomas of the trunk (n = 154), of the head and neck (HN; n = 76), and lentigo maligna (LM) and lentigo maligna melanoma (LMM; for both LM and LMM, n = 76). Data were collected on school-age sun exposure and occupational and recreational sun exposure in adulthood. Odds ratios (OR) and 95% CIs were calculated using polytomous logistic regression. RESULTS: HN melanoma patients were substantially more likely than trunk patients to have higher levels of sun exposure in adulthood (OR, 2.43; 95% CI, 0.98 to 5.99) and specifically, higher levels of occupational exposure (OR, 3.25; 95% CI, 1.32 to 8.00), but lower levels of recreational sun exposure (OR, 0.50; 95% CI, 0.21 to 1.19). LM and LMM patients reported higher occupational exposure and lower recreational sun exposure than trunk melanoma patients, although this was not significant. We found no significant differences between the groups for school-age sun exposures. CONCLUSION: Melanomas developing at different body sites are associated with distinct patterns of sun exposure. Melanomas of the head and neck are associated with chronic patterns of sun exposure whereas trunk melanomas are associated with intermittent patterns of sun exposure, supporting the hypothesis that melanomas may arise through divergent causal pathways.
Several decades of research have confirmed that the relationship between sun exposure and melanoma is complex and is strongly modified by host factors including pigmentation and propensity to develop nevi.1 The pattern of exposure to sunlight has been hypothesized to be the key determinant in melanoma development, with intermittent exposure considered to be causal and chronic exposure protective.2 While epidemiologic evidence has accumulated in support of the intermittent exposure hypothesis,3 several features of the occurrence of melanoma indicate that chronic exposure to sunlight has a causal role. We have previously proposed a divergent pathway model for cutaneous melanoma irrespective of histologic classification which seeks to explain the development of melanoma by incorporating the role of sunlight, host susceptibility, and anatomic site.4,5 In support of this model, we found marked differences in the prevalence of nevi and solar keratoses between patients with head and neck melanomas and those with trunk melanomas.5 Similar findings have been reported by others,6,7 and coupled with emerging molecular evidence that somatic genetic aberrations in cutaneous melanoma also vary by site,8,9 suggest that melanomas arising at different anatomic sites may have different causal pathways.10 Here, we further explore the causation of melanoma by sun exposure. Our primary hypothesis was that people with melanomas of the head and neck would be exposed to greater cumulative doses of sunlight than those with melanomas of the trunk.
Methodologic details of patient selection and data collection have been described previously.5 Briefly, we used a case-case study design to test our hypothesis, in which one group of melanoma patients served as the reference group to which two other groups of melanoma patients were compared. Approval to undertake the study was granted by the human research ethics committee of the Queensland Institute of Medical Research (Brisbane, Queensland, Australia) and all participants gave their written informed consent.
Patients
Data Collection A research nurse unaware of the study hypotheses checked questionnaires for completeness and counted solar keratoses and nevi using standard definitions as previously described.5 Histologic details about the type and thickness of melanoma were abstracted from pathology reports by a dermatologist who was not otherwise involved in data collection.
Statistical Analysis To estimate sun exposure to the trunk during each employment period, we weighted the measures of ambient exposure to account for the proportion of time that the trunk was not covered by clothing on work days and nonwork days. People who reported never wearing clothing covering the trunk during a given time period were accorded a weighting factor of 1 (that is, trunk skin was exposed to sunlight all of the time). Those reporting rarely, sometimes, often, or always wearing clothing covering the trunk scored 0.8, 0.5, 0.2, and 0, respectively.
We measured the strength of association between categories of sun exposure and each melanoma group by calculating the odds ratio (OR) and 95% CIs with multivariable logistic regression analysis. Patients with trunk melanomas comprised the control group, to which the two other groups were compared. We adjusted for Breslow thickness, sex, and exact age in years and also included a term for age-squared in all models to adjust for residual confounding by nonlinear effects of age. Analyses were initially conducted using data from all melanoma patients and then repeated restricted to patients with invasive melanomas only. For logistic regression analysis, we categorized sun exposure indices at tertile cut points, and the group with lowest cumulative sun exposure served as the reference category. We fitted additional models including terms for pigmentation factors and numbers of nevi as potential confounders. To test for trend, categoric data were included in the model as continuous variables, and the corresponding Wald statistic was compared to a
Role of Funding Organizations
Of the 498 patients meeting the eligibility criteria, written consent was received from treating doctors to approach 452 patients (91%). Of these, 328 patients (73%) gave written consent to participate (65 refusals and 59 nonresponders—additional follow-up of nonresponders was not permitted by the cancer registry); 316 patients returned a completed questionnaire, and 306 patients also completed the physical examination (trunk melanoma, n = 154; head and neck melanoma, n = 76; LM and LMM n = 76.) Characteristics of study participants are outlined in Table 1.
Sun Exposure at School Ages Measures of sun exposure in elementary school and high school, and measures of association with each melanoma type, are presented in Table 2. No consistent differences in self-reported measures of sun exposure at elementary school were observed among the three groups of melanoma patients. There was some evidence that the frequency of severe sunburns during elementary school years was higher among the group of patients with LM or LMM, however this was not significant.
Overall, study participants reported lower levels of outdoor exposure during high school ages compared with elementary school ages. Nevertheless, more than 80% reported at least one episode of blistering or peeling sunburn during the high school years, and almost half reported more than five such episodes (Table 2). No significant associations with outdoor exposures during high school years were observed.
Ambient Sun Exposure During Adulthood
We partitioned total cumulative sun exposure into occupational and recreational components and noted considerable differences in risk. Patients with melanomas of the head and neck reported significantly higher levels of ambient exposure on work days (OR, 3.25; 95% CI, 1.32 to 8.00) than patients with trunk melanomas, but were significantly less likely to be exposed to the sun on nonwork days (medium exposure nonwork OR, 0.38; 95% CI, 0.17 to 0.83; high exposure nonwork OR, 0.50; 95% CI, 0.21 to 1.19; Table 3). Similar patterns of association were observed in the analyses restricted to invasive melanomas. Patients with LM and LMM were also found to be less likely than those with trunk melanoma to have high levels of recreational sun exposure, although this was not significant. Compared with those with melanomas of the trunk, people with head and neck melanoma were significantly more likely to receive more than half of their total dose of sunlight through workday exposures (Table 3). We fitted further models with terms for skin type and nevus number but these made little difference to the risk estimates (not shown). We compared solar keratosis counts for each of the three groups of melanoma patients according to their occupational sun exposure to assess whether the self-reported sun exposure histories correlated with objective findings of solar damage (Fig 1). Those with high levels of occupational sun exposure had higher solar keratosis counts than people with low levels of exposure in all groups of melanoma patients.
Sun Exposure to the Trunk During Adulthood A history of sun exposure to the trunk was not uncommon in this sample of melanoma patients. Total duration of sun exposure to the trunk ranged from 0 hours to 44,394 hours (median, 1,732 hours; lower tertile, 564 hours; upper tertile, 3,339 hours). Aside from a small number of patients who were outdoor workers, most participants acquired their trunk sun exposure on nonwork days (median sun exposure to the trunk: occupational 0 hours, recreational 918 hours). Patients with melanomas of head and neck typically reported similar levels of trunk sun exposure to people with melanomas of the trunk on working days, but reported lower levels of trunk sun exposure on nonwork days (Table 3). Associations of similar direction and magnitude were observed for patients with invasive melanomas of the head and neck and for patients with LM and LMM.
Speculation that melanomas at different anatomic sites may arise through different causal pathways is not new,11-13 and while empirical evidence in support of this notion has been reported,4,6,14,15 we are not aware of previous analytic studies designed to test this hypothesis. We have found that cutaneous melanomas developing on the head and neck were significantly more likely to occur in people with high levels of total sun exposure during adulthood, whereas melanomas on the trunk tended to occur on people with lower levels of sunlight exposure overall, but who reported higher levels of recreational exposure on the chest and back. We found no consistent evidence that patterns of sun exposure during school ages differed for melanomas arising on different body sites. Our findings accord with earlier studies that reported that outdoor workers have a relative excess of head and neck melanomas and a deficit of melanomas at other body sites.16-19 It has been inferred from such data that melanomas arising on the head and neck are associated with chronic exposure to high levels of sunlight, however a spurious association cannot be excluded since these earlier studies did not exclude patients with LMM, the histologic subtype with a predilection for the face and widely acknowledged to be caused by chronic sun exposure.20 Because we separately analyzed patients with LMM from those with SSM and NM of the head and neck, the observed differences in sun exposure by anatomic site cannot be explained by this potential bias. We considered the possibility that our findings were due simply to the head and neck melanoma patients being older than the patients with melanoma of the trunk, and thus had greater opportunity for accumulating high levels of sun exposure. We addressed this issue firstly by oversampling younger patients with head and neck melanoma and then adjusting for the linear and nonlinear effects of age. Confounding by other factors was also considered. We included terms for skin type in the models and this made no substantial difference to the associations. Adjusting for nevus number when considering the association between sun exposure and melanoma is contentious, with arguments for and against this practice centered on the biologic role of nevi in the causal pathway to melanoma.21We have presented risk estimates that were not adjusted for numbers of nevi, although including a term for nevus number in the various models made almost no difference to the estimates of risk. Recall bias is an improbable explanation for the associations with overall measures of ambient sun exposure observed here, since all participants in the study were case patients with melanoma who were not aware of the specific hypotheses being tested. While some degree of misclassification is inevitable when measuring sun exposure,22 it is unlikely that this would be differential with respect to the site of melanoma. Similarly, histological misclassification of melanomas (eg, LMM with SSM), while possible, cannot explain the discordant patterns of risk by anatomic site. Errors of inference might arise if in situ melanomas differ in their causation from invasive melanomas, although we are not aware of any data to suggest this effect and analyses restricted to patients with invasive melanomas only made essentially no difference to most estimates of risk. Assuming these findings reflect real differences in the association between sunlight and melanoma at different anatomic sites, then the mechanisms remain to be defined. The prevailing paradigm to explain melanoma development is the intermittent exposure hypothesis, which proposes that episodic exposure of pigment cells to sunlight is primarily responsible for initiating tumorigenesis.21,23,24 Numerous epidemiologic studies have been cited as providing evidence to support this model,3,25 and our data might also be interpreted in this way, at least in part, since we found that people with trunk melanomas were significantly more likely to report high levels of recreational sun exposure than people with head and neck melanomas. However, the strongly positive association of head and neck melanomas with occupational sun exposure that we observed argues against melanomas at these sites arising through an intermittent exposure pathway. How might these observations be reconciled? One explanation is that melanocytes at different anatomic locations have different propensities to undergo malignant change,26 with melanocytes on the trunk postulated to be more prone to neoplasia than those on the head and neck. Such differences might be innate, due to anatomically-determined properties of melanocytes13,27,28 or may reflect an acquired characteristic reflecting the pattern of exposure at the site in question.24 Another possibility, and not withstanding either of the preceding explanations, is that the site at which a melanoma develops reflects the susceptibility of host melanocytes to proliferate in response to sunlight. Under this hypothesis, people with an inherently low propensity to develop nevi require high levels of sunlight to develop melanoma and their tumors will tend to arise on habitually exposed sites such as the head and neck. Data in support of this hypothesis have been published.6,7 In summary, we have found evidence that melanomas at different anatomic sites have different associations with patterns of sun exposure. These findings, together with observed differences in gene expression between melanomas8,9and the distinct age distributions for melanomas of the trunk and head15,29 suggest that melanomas can arise through different causal pathways. Pursuing this line of enquiry through molecular and histologic approaches may yield new insights into the underlying mechanisms of this phenomenon.
The authors indicated no potential conflicts of interest.
Conception and design: David C. Whiteman, Adèle C. Green Financial support: David C. Whiteman Collection and assembly of data: David C. Whiteman, Peter Watt, Marcia B. Davis, Adèle C. Green Data analysis and interpretation: David C. Whiteman, Mark Stickley, Peter Watt, Maria Celia Hughes, Adèle C. Green Manuscript writing: David C. Whiteman, Maria Celia Hughes, Adèle C. Green Final approval of manuscript: David C. Whiteman, Mark Stickley, Peter Watt, Maria Celia Hughes, Marcia B. Davis, Adèle C. Green
Supported by grants from the Queensland Cancer Fund and the National Cancer Institute (CA 88363-01A1). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute. Presented in part during an invited plenary session at the 6th World Congress on Melanoma, Vancouver, British Columbia, Canada, September 8, 2005 (D.C.W.). Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
1. Armstrong BK, Kricker A: The epidemiology of UV induced skin cancer. J Photochem Photobiol B 63:8-18, 2001[CrossRef][Medline] 2. Armstrong BK: How sun exposure causes skin cancer: An epidemiological perspective, in Hill D, Elwood JM, English DR (eds): Prevention of Skin Cancer: Cancer Prevention, Cancer Causes. Boston, MA, Kluwer Academic Publishers, 2004 3. Elwood JM, Jopson J: Melanoma and sun exposure: An overview of published studies. Int J Cancer 73:198-203, 1997[CrossRef][Medline] 4. Whiteman DC, Green A, Parsons PG: p53 Expression and risk factors for cutaneous melanoma: A case-control study. Int J Cancer 77:843-848, 1998[CrossRef][Medline] 5. Whiteman DC, Watt P, Purdie DM, et al: Melanocytic nevi, solar keratoses, and divergent pathways to cutaneous melanoma. J Natl Cancer Inst 95:806-812, 2003 6. Bataille V, Sasieni P, Grulich A, et al: Solar keratoses: A risk factor for melanoma but negative association with melanocytic naevi. Int J Cancer 78:8-12, 1998[CrossRef][Medline] 7. Carli P, Palli D: Re: Melanocytic nevi, solar keratoses, and divergent pathways to cutaneous melanoma. J Natl Cancer Inst 95:1801, 2003 8. Maldonado JL, Fridlyand J, Patel H, et al: Determinants of BRAF mutations in primary melanomas. J Natl Cancer Inst 95:1878-1890, 2003 9. Curtin JA, Fridlyand J, Kageshita T, et al: Distinct sets of genetic alterations in melanoma. N Engl J Med 353:2135-2147, 2005 10. Rivers JK: Is there more than one road to melanoma? Lancet 363:728-730, 2004[CrossRef][Medline] 11. Houghton A, Flannery J, Viola MV: Malignant melanoma in Connecticut and Denmark. Int J Cancer 25:95-104, 1980[Medline] 12. Ackerman AB: Malignant melanoma: A unifying concept. Hum Pathol 11:591-595, 1980[Medline] 13. Green A: A theory of site distribution of melanomas: Queensland, Australia. Cancer Causes Control 3:513-516, 1992[CrossRef][Medline] 14. Carli P, Massi D, Santucci M, et al: Cutaneous melanoma histologically associated with a nevus and melanoma de novo have a different profile of risk: Results from a case-control study. J Am Acad Dermatol 40:549-557, 1999[CrossRef][Medline] 15. Elwood JM, Gallagher RP: Body site distribution of cutaneous malignant melanoma in relationship to patterns of sun exposure. Int J Cancer 78:276-280, 1998[CrossRef][Medline] 16. Beral V, Robinson N: The relationship of malignant melanoma, basal and squamous skin cancers to indoor and outdoor work. Br J Cancer 44:886-891, 1981[Medline] 17. Vagero D, Ringback G, Kiviranta H: Melanoma and other tumours of the skin among office, other indoor workers and outdoor workers in Sweden 1961-1979. Br J Cancer 53:507-512, 1986[Medline] 18. Linet MS, Malker HS, Chow WH, et al: Occupational risks for cutaneous melanoma among men in Sweden. J Occup Environ Med 37:1127-1135, 1995[CrossRef][Medline] 19. Hakansson N, Floderus B, Gustavsson P, et al: Occupational sunlight exposure and cancer incidence among Swedish construction workers. Epidemiology 12:552-557, 2001[CrossRef][Medline] 20. Cohen LM: Lentigo maligna and lentigo maligna melanoma. J Am Acad Dermatol 33:923-936, 1995[CrossRef][Medline] 21. Armstrong BK: Epidemiology of malignant melanoma: Intermittent or total accumulated exposure to the sun. J Dermatol Surg Oncol 14:835-849, 1988[Medline] 22. English DR, Armstrong BK, Kricker A: Reproducibility of reported measurements of sun exposure in a case-control study. Cancer Epidemiol Biomarkers Prev 7:857-863, 1998[Abstract] 23. Holman CD, Armstrong BK, Heenan PJ: A theory of the etiology and pathogenesis of human cutaneous malignant melanoma. J Natl Cancer Inst 71:651-656, 1983[Medline] 24. Gilchrest BA, Eller MS, Geller AC, et al: The pathogenesis of melanoma induced by ultraviolet radiation. N Engl J Med 340:1341-1348, 1999 25. Gandini S, Sera F, Cattaruzza MS, et al: Meta-analysis of risk factors for cutaneous melanoma: II. Sun exposure. Eur J Cancer 41:45-60, 2005[CrossRef][Medline] 26. Clark WH, Elder DE, Guerry D, et al: A study of tumor progression: The precursor lesions of superficial spreading and nodular melanoma. Hum Pathol 15:1147-1165, 1984[Medline] 27. Abdel-Malek ZA, Swope VB, Nordlund JJ, et al: Proliferation and propagation of human melanocytes in vitro are affected by donor age and anatomical site. Pigment Cell Res 7:116-122, 1994[CrossRef][Medline] 28. Silvers WK, Mintz B: Differences in latency and inducibility of mouse skin melanomas depending on the age and anatomic site of the skin. Cancer Res 58:630-632, 1998 29. Bulliard J-L: Site-specific risk of cutaneous malignant melanoma and pattern of sun exposure in New Zealand. Int J Cancer 85:627-632, 2000[CrossRef][Medline] Submitted February 9, 2006; accepted April 21, 2006. This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2006 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
50 years and > 50 years) and sampled with the aim of achieving equal numbers in each. 
2 with one degree of freedom. Statistical significance was determined at 
= .05, and all tests for statistical significance were two sided. All analyses were performed in SAS, version 9.1 (SAS Institute, Cary, NC). 
25th centile; 
median; 
