The aggregation of early-onset melanoma in young Western Australian families. Issue 3 (June 2015)

• Record Type: Journal Article
• Authors: Ward, S.V.; Dowty, J.G.; Webster, R.J.; Cadby, G.; Glasson, E.J.; Heyworth, J.S.; Emery, J.; Cole, J.M.; Millward, M.J.; Wood, F.M.; Palmer, L.J.
• Is Part Of: Cancer epidemiology. Volume 39:Issue 3(2015)
• Page Start: 346
• Page End: 352
• Publication Date: 2015-06
• Subjects: Melanoma -- Cancer -- Hereditary cancer syndromes -- Data linkage -- Cox proportional hazards model; Cancer -- Epidemiology -- Periodicals; Cancer -- Prevention -- Periodicals; Cancer -- Diagnosis -- Periodicals; Carcinogenesis -- Periodicals; 616.994005
• DOI: 10.1016/j.canep.2015.03.007
• Languages: English
• ISSNs: 1877-7821

Abstract

Highlights: Cox proportional hazards models investigated cancer aggregation in young families. Melanoma incidence was significantly greater in relatives of melanoma cases. The observed hazard ratio likely reflects early-age at onset in melanoma cases. No significant evidence was found for co-aggregation of other cancers with melanoma. Abstract: Background: Few studies have examined the familial aggregation of melanoma or its co-aggregation with other cancers using whole-population based designs. This study aimed to investigate aggregation patterns in young Western Australian families, using population-based linked health data to identify individuals born in Western Australia between 1974 and 2007, their known relatives, and all incident cancer diagnoses within the resulting 1, 506, 961 individuals. Methods: Cox proportional hazards regression models were used to compare the risk of melanoma for first-degree relatives of melanoma cases to that for first-degree relatives of controls, with bootstrapping used to account for correlations within families. The risk of (i) developing melanoma based on the number of first-degree relatives with other cancers, and (ii) developing non-melanoma cancers based on the number of first-degree relatives diagnosed with melanoma was also investigated. Results: First-degree relatives of melanoma cases had a significantly greater incidence of melanoma than first-degree relatives of individuals not affected with melanoma (Hazard Ratio (HR) = 3.58, 95% bootstrap confidence interval (CI): 2.43–5.43). Sensitivity analyses produced a higher hazard ratio estimate when restricted to melanoma cases diagnosed before 40 years of age (HR = 3.77, bootstrap 95% CI: 2.49–6.39) and a lower estimate when only later-onset cases (>40 years) were considered (HR = 2.45, bootstrap 95% CI: 1.23–4.82). No significant evidence was found for co-aggregation between melanoma and any other cancers. Conclusions: Results indicated a strong familial basis of melanoma, with the higher than expected hazard ratio observed likely to reflect early-age at onset cases in this young cohort, supported by the results of the sensitivity analyses. Exploratory analyses suggested that the determinants of melanoma causing the observed aggregation within families may be independent of other malignancies, although these analyses were limited by the young age of the sample. Determining familial aggregation patterns will provide valuable knowledge regarding improved clinical risk prediction and the underlying biological mechanisms of melanoma and other cancers.