Risk of Death Due to Melanoma and Other Causes in Patients With Thin Cutaneous Melanomas.

Abstract

Importance: Most patients who present with primary cutaneous melanomas have thin tumors (≤1.0 mm in Breslow thickness, ie, pT1a and pT1b). Although their prognosis is generally considered to be excellent, there is limited precise information on the association of risk of death with specific Breslow measurements in thin lesions.

Objective: To assess the relative effect of a 0.8-mm Breslow thickness threshold with respect to the incidence of both melanoma-related and nonmelanoma-related death.

Design, setting, and participants: Registry data for all Australians diagnosed with thin invasive primary melanomas between 1982 and 2014 were analyzed. Data were extracted from all 8 Australian state and territory population-based cancer registries. Dates and causes of death were obtained from the Australian National Death Index. Adults diagnosed with a first invasive primary melanoma of 1.0 mm or smaller in thickness were included.

Exposure: First invasive primary melanoma between 1982 and 2014.

Main outcomes and measures: The primary outcomes were melanoma-related deaths and nonmelanoma-related deaths. Competing-risk regression analyses and cause-specific analyses were performed to investigate the relationships between Breslow thickness subcategory (<0.8 mm versus ≥0.8 mm by 0.1-mm increments) and the primary outcomes.

Results: Overall, a cohort of 144 447 participants was included. The median (range) age was 56 (18-101) years and 78 014 (54.0%) were men. Median (IQR) follow-up was 15.0 (9.5-23.3) years. Crude incidence rates of melanoma-related death 20 years after diagnosis were 6.3% (95% CI, 6.1%-6.5%) for the whole cohort, 6.0% (95% CI, 5.7%-6.2%) for tumors smaller than 0.8 mm, and 12.0% (95% CI, 11.4%-12.6%) for tumors 0.8 to 1.0 mm. The corresponding 20-year melanoma-specific survival rates were 91.9% (95% CI, 91.6%-92.1%), 94.2% (95% CI, 94.0%-94.4%), and 87.8% (95% CI, 87.3%-88.3%), respectively. On multivariable analysis, tumor thickness of 0.8 to 1.0 mm was significantly associated with both a greater absolute risk of melanoma-related death (subdistribution hazard ratio, 2.92; 95% CI, 2.74-3.12) and a greater rate of melanoma-related death (hazard ratio, 2.98; 95% CI, 2.79-3.18) than thinner tumors (<0.8 mm). Risk of death from nonmelanoma-related causes was not associated with Breslow thickness.

Conclusions and relevance: In this study, the risk of melanoma-related death increased significantly for patients with primary tumors of 0.8 to 1.0 mm in thickness. The risk of death from nonmelanoma-ralated causes was similar across Breslow thicknesses of 0.1 to 1.0 mm. This analysis suggests that a 0.8-mm threshold for guiding the care of patients with thin primary melanomas.