Implications of age and conditional survival estimates for patients with melanoma

Abstract

Overall cancer incidence is decreasing, whereas melanoma cases are increasing. Conditional survival estimates offer a more accurate prognosis for patients the farther they are from time of diagnosis. The effect of age and stage on a melanoma patient's conditional survival estimate is unknown. Surveillance, Epidemiology, and End Results data were utilized to identify newly diagnosed cutaneous melanoma patients (N=95 041), from 1998 to 2005, with up to 12 years of follow-up. Estimates of disease-specific survival by stage and age were determined by Cox regression analysis and transformed to estimated conditional 5-year survival. Localized melanoma patients have an excellent 5-year survival at diagnosis and over subsequent years. For patients with localized and regional disease, an age effect is present for disease-specific mortality when comparing older patients (70-79 years) with younger patients (<30 years): hazard ratio (HR) for mortality 3.79 [95% confidence interval (CI) 3.01-4.84] and HR 2.36 (95% CI 1.93-2.91), respectively. No age effect difference is observed in disease-specific survival for advanced disease: HR 1.14 (95% CI 0.87-1.53). Over time, conditional survival estimates improve for older patients with localized and regional disease. This improvement is not seen in distant disease, neither is the age gradient. Disease-specific mortality and conditional survival for patients with localized and regional melanomas are initially impacted by older age, with effects dissipating over time. Age does not affect survival in patients with advanced disease. Understanding the conditional 5-year disease-specific survival of melanoma based on age and stage can help patients and physicians, informing decision-making about treatment and surveillance.