RESUMO
BACKGROUND: Melanoma incidence rates have increased dramatically in white populations worldwide during the past several decades. A more modest increase has been observed for melanoma-related mortality. Cause-specific and disease-free survivals are related to tumor characteristics, gender, age, and possible anatomic site. It is difficult to accurately assess these trends without information on tumor thickness that is often unavailable. OBJECTIVE: This study determines trends in melanoma incidence, mortality, and survival in Saskatchewan for a 30-year period, incorporating analysis of tumor thickness. METHODS: Information about cases of primary cutaneous melanoma for the 30-year period 1970-1999 was obtained from the population-based Saskatchewan Cancer Registry. A 50% random sample of charts was reviewed to collect information about Breslow depth, Clark level, and other demographic data not available from the Registry. Multivariate regression analysis was used to determine the significance of prognostic factors on incidence and five-year relative survival rates. RESULTS: The number of patients registered increased dramatically during the study period. The increase was greatest for thin lesions in all age groups. Anatomic site varied by gender. Head and neck tumors showed continual increase in risk with increasing age. Mortality rates in females have been stable over time but increased for males in the 1990s. The prognostic factors tht predicted excess mortality at five years were tumor thickness, Clark level, and gender. CONCLUSION: The observed increase in melanoma appears to be real and not the result of increased surveillance or screening. Tumor characteristic (Breslow depth, Clark level) and gender were significant prognostic indicators of five-year excess mortality.