Competing risk analysis of mortality from invasive cutaneous melanoma in New South Wales: a population‐based study, 1988–2007

Abstract

Abstract Objective: Accurate estimates of risk of death from melanoma, based on the most recent information, are desirable, especially if secular improvements in survival have occurred. This study aims to investigate prognostic factors and temporal changes in mortality from primary invasive cutaneous melanoma (CM) and to predict cumulative probabilities of death from CM. Methods: Cases of CM from the NSW Central Cancer Registry (NSWCCR) diagnosed in 1988–2007 were analysed. We used Fine and Gray competing risks models to investigate prognostic factors associated with CM mortality, along with period effects of year of diagnosis. Adjusted cumulative probabilities of CM death were then estimated. Results: Of 52,330 CM cases, 5291 (10%) died from CM and 8290 (16%) from other causes. Patients with tumours thicker than 4 mm had 9.5 times the risk of death from CM compared to those with tumours 1 mm or less (subhazard ratio [SHR] 9.52; 95%CI:8.42–10.77). Risk of melanoma death was 31% lower in 2003–2007 compared to 1988–1992 (SHR 0.69; 95%CI: 0.63–0.76). Other risk factors for CM mortality included older age and male gender. Assuming the estimated period effect for a diagnosis in 2003–2007 applies now, the predicted probability of CM death within 10 years of diagnosis of a tumour 4+ mm thick is 26% in males and 19% in females. Conclusion: This study highlights the importance of awareness and early detection and shows a significant improvement in survival from CM since 1988.

Keywords

• cutaneous melanoma
• population‐based
• mortality
• tumour thickness
• competing risks